Marino - The ICU Book - 3a Edition

1,352 Pages • 305,899 Words • PDF • 11 MB
Uploaded at 2021-09-24 08:35

This document was submitted by our user and they confirm that they have the consent to share it. Assuming that you are writer or own the copyright of this document, report to us by using this DMCA report button.

Courtesy of


Authors: Marino, Paul L. Title: ICU Book, The, 3rd Edition Copyright ©2007 Lippincott Williams & Wilkins ISBN: 0-7817-4802-X

Table of Contents Section I - Basic Science Review Basic Science Review Chapter 1 - Circulatory Blood Flow Chapter 2 - Oxygen and Carbon Dioxide Transport Section II - Preventive Practices in the Critically Ill Preventive Practices in the Critically Ill Chapter 3 - Infection Control in the ICU Chapter 4 - Alimentary Prophylaxis Chapter 5 - Venous Thromboembolism Section III - Vascular Access Vascular Access Chapter 6 - Establishing Venous Access Chapter 7 - The Indwelling Vascular Catheter Section IV - Hemodynamic Monitoring Hemodynamic Monitoring Chapter 8 - Arterial Blood Pressure Chapter 9 - The Pulmonary Artery Catheter Chapter 10 - Central Venous Pressure and Wedge Pressure Chapter 11 - Tissue Oxygenation Section V - Disorders of Circulatory Flow Disorders of Circulatory Flow Chapter 12 - Hemorrhage and Hypovolemia Chapter 13 - Colloid and Crystalloid Resuscitation Chapter 14 - Acute Heart Failure Syndromes

Chapter 15 - Cardiac Arrest Chapter 16 - Hemodynamic Drug Infusions Section VI - Critical Care Cardiology Critical Care Cardiology Chapter 17 - Early Management of Acute Coronary Syndromes Chapter 18 - Tachyarrhythmias Section VII - Acute Respiratory Failure Acute Respiratory Failure Chapter 19 - Hypoxemia and Hypercapnia Chapter 20 - Oximetry and Capnography Chapter 21 - Oxygen Inhalation Therapy Chapter 22 - Acute Respiratory Distress Syndrome Chapter 23 - Severe Airflow Obstruction Section VIII - Mechanical Ventilation Mechanical Ventilation Chapter 24 - Principles of Mechanical Ventilation Chapter 25 - Modes of Assisted Ventilation Chapter 26 - The Ventilator-Dependent Patient Chapter 27 - Discontinuing Mechanical Ventilation Section IX - Acid-Base Disorders Acid-Base Disorders Chapter 28 - Acid-Base Interpretations Chapter 29 - Organic Acidoses Chapter 30 - Metabolic Alkalosis Section X - Renal and Electrolyte Disorders Renal and Electrolyte Disorders Chapter 31 - Oliguria and Acute Renal Failure

Chapter 32 - Hypertonic and Hypotonic Conditions Chapter 33 - Potassium Chapter 34 - Magnesium Chapter 35 - Calcium and Phosphorus Section XI - Transfusion Practices in Critical Care Transfusion Practices in Critical Care Chapter 36 - Anemia and Red Blood Cell Transfusions in the ICU Chapter 37 - Platelets in Critical Illness Section XII - Disorders of Body Temperature Disorders of Body Temperature Chapter 38 - Hyperthermia and Hypothermia Syndromes Chapter 39 - Fever in the ICU Section XIII - Inflammation and Infection in the ICU Inflammation and Infection in the ICU Chapter 40 - Infection, Inflammation, and Multiorgan Injury Chapter 41 - Pneumonia in the ICU Chapter 42 - Sepsis from the Abdomen and Pelvis Chapter 43 - The Immuno-Compromised Patient Chapter 44 - Antimicrobial Therapy Section XIV - Nutrition and Metabolism Nutrition and Metabolism Chapter 45 - Metabolic Substrate Requirements Chapter 46 - Enteral Tube Feeding Chapter 47 - Parenteral Nutrition Chapter 48 - Adrenal and Thyroid Dysfunction Section XV - Critical Care Neurology Critical Care Neurology

Chapter 49 - Analgesia and Sedation Chapter 50 - Disorders of Mentation Chapter 51 - Disorders of Movement Chapter 52 - Stroke and Related Disorders Section XVI - Toxic Ingestions Toxic Ingestions Chapter 53 - Pharmaceutical Toxins & Antidotes Section XVII: Appendices Appendix 1 - Units and Conversions Appendix 2 - Selected Reference Ranges Appendix 3 - Clinical Scoring Systems

Author Paul L. Marino MD, PhD, FCCM Physician-in-Chief Saint Vincent's Midtown Hospital, New York, New York; Clinical Associate Professor, New York Medical College, Valhalla, New York

Contributor Kenneth M. Sutin MD, FCCM Dr. Kenneth Sutin contributed to the final 13 chapters of this book Department of Anesthesiology, Bellevue Hospital Center; Associate Professor of Anesthesiology & Surgery, New York University School of Medicine, New York, New York

Illustrator Patricia Gast

Secondary Editors Brian Brown Acquisitions Editor Nicole Dernoski Managing Editor Tanya Lazar Managing Editor Bridgett Dougherty Production Manager Benjamin Rivera Senior Manufacturing Manager Angela Panetta Marketing Manager Doug Smock Creative Director Nesbitt Graphics, Inc. Production Services RR Donnelley

Printer Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Dedication To Daniel Joseph Marino, My 18-year-old son. No longer a boy, And not yet a man, But always terrific. Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Quote I would especially commend the physician who, in acute diseases, by which the bulk of mankind are cutoff, conducts the treatment better than others. —HIPPOCRATES Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Preface to Third Edition The third edition of The ICU Book marks its 15 th year as a fundamental sourcebook in critical care. This edition continues the original intent to provide a generic textbook that presents fundamental concepts and patient care practices that can be used in any intensive care unit, regardless of the specialty focus of the unit. Highly specialized areas, such as obstetrical emergencies, thermal injury, and neurocritical care, are left to more qualified authors and their specialty textbooks. Most of the chapters in this edition have been completely rewritten (including 198 new illustrations and 178 new tables), and there are two new chapters on infection control in the ICU (Chapter 3) and disorders of temperature regulation (Chapter 38). Most chapters also include a final section (called A Final Word) that contains an important take-home message from the chapter. The references have been extensively updated, with emphasis on recent reviews and clinical practice guidelines. The ICU Book has been unique in that it reflects the voice of one author. This edition welcomes the voice of another, Dr. Kenneth Sutin, who added his expertise to the final 13 chapters of the book. Ken and I are old friends who share the same view of critical care medicine, and his contributions add a robust quality to the material without changing the basic personality of the work. Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Preface to First Edition In recent years, the trend has been away from a unified approach to critical illness, as the specialty of critical care becomes a hyphenated attachment for other specialties to use as a territorial signpost. The landlord system has created a disorganized array of intensive care units (10 different varieties at last count), each acting with little communion. However, the daily concerns in each intensive care unit are remarkably similar because serious illness has no landlord. The purpose of The ICU Book is to present this common ground in critical care and to focus on the fundamental principles of critical illness rather than the specific interests for each intensive care unit. As the title indicates, this is a ‘generic’ text for all intensive care units, regardless of the name on the door. The present text differs from others in the field in that it is neither panoramic in scope nor overly indulgent in any one area. Much of the information originates from a decade of practice in intensive care units, the last three years in both a Medical ICU and a Surgical ICU. Daily rounds with both surgical and medical housestaff have provided the foundation for the concept of generic critical care that is the theme of this book. As indicated in the chapter headings, this text is problem-oriented rather than disease-oriented, and each problem is presented through the eyes of the ICU physician. Instead of a chapter on GI bleeding, there is a chapter of the principles of volume resuscitation and two others on resuscitation fluids. This mimics the actual role of the ICU physician in GI bleeding, which is to manage the hemorrhage. The other features of the problem such as locating the bleeding site, are the tasks of other specialists. This is how the ICU operates and this is the specialty of critical care. Highly specialized topics such as burns, head trauma, and obstetric emergencies are not covered in this text. These are distinct subspecialties with their own texts and their own experts, and devoting a few pages to each would merely complete and outline rather than instruct. The emphasis on fundamentals in The ICU Book is meant not only as a foundation for patient care but also to develop a strong base in clinical problem solving for any area of medicine. There is a tendency to rush past the basics in the stampede to finish formal training, and this leads to empiricism and irrational practice habits. Why a fever should or should not be treated, or whether a blood pressure cuff provides accurate readings, are questions that must be dissected carefully in the early stages of training, to develop the reasoning skills needed to be effective in clinical problems solving. This inquisitive stare must replace the knee-jerk approach to clinical problems if medicine is to advance. The ICU Book helps to develop this stare. Wisely or not, the use of a single author was guided by the desire to present a uniform view. Much of the information is accompanied by published works listed at the end of each chapter and anecdotal tales are held to a minimum. Within an endeavor such as this, several shortcomings are inevitable, some omissions are likely and bias may occasionally replace sound judgment. The hope is that these deficiencies are few. Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Acknowledgments Acknowledgements are few but well deserved. First to Patricia Gast, the illustrator for this edition, who was involved in every facet of this work, and who added an energy and intelligence that goes well beyond the contributions of medical illustrators. Also to Tanya Lazar and Nicole Dernoski, my editors, for understanding the enormous time committment required to complete a work of this kind. And finally to the members of the executive and medical staff of my hospital, as well as my personal staff, who allowed me the time and intellectual space to complete this work unencumbered by the daily (and sometimes hourly) tasks involved in keeping the doors of a hospital open. Copyright (c) 2000-2006 Ovid Technologies, Inc. Version: rel10.3.2, SourceID 1.12052.1.159

Basic Science Review The first step in applying the scientific method consists in being curious about the world. --Linus Pauling

Chapter 1 Circulatory Blood Flow When is a piece of matter said to be alive? When it goes on “doing something,” moving, exchanging material with its environment. --Erwin Schrodinger The human organism has an estimated 100 trillion cells that must go on exchanging material with the external environment to stay alive. This exchange is made possible by a circulatory system that uses a muscular pump (the heart), an exchange fluid (blood), and a network of conduits (blood vessels). Each day, the human heart pumps about 8,000 liters of blood through a vascular network that stretches more than 60,000 miles (more than twice the circumference of the Earth!) to maintain cellular exchange (1). This chapter describes the forces responsible for the flow of blood though the human circulatory system. The first half is devoted to the determinants of cardiac output, and the second half describes the forces that influence peripheral blood flow. Most of the concepts in this chapter are old friends from the physiology classroom.

Cardiac Output Circulatory flow originates in the muscular contractions of the heart. Since blood is an incompressible fluid that flows through a closed hydraulic loop, the volume of blood ejected by the left side of the heart must equal the volume of blood returning to the right side of the heart (over a given time period). This conservation of mass (volume) in a closed hydraulic system is known as the principle of continuity (2), and it indicates that the stroke output of the heart is the principal determinant of circulatory blood flow. The forces that govern cardiac stroke output are identified in Table 1.1. TABLE 1.1 The Forces that Determine Cardiac Stroke Output



Clinical Parameters


The load imposed on resting muscle that stretches the muscle to a new length

End-diastolic pressure


The velocity of muscle contraction when muscle load is fixed

Cardiac stroke volume when preload and afterload are constant


The total load that must be moved by a muscle when it contracts

Pulmonary and systemic vascular resistances


Preload If one end of a muscle fiber is suspended from a rigid strut and a weight is attached to the other free end, the added weight will stretch the muscle to a new length. The added weight in this situation represents a force called the preload, which is a force imposed on a resting muscle (prior to the onset of muscle contraction) that stretches the muscle to a new length. According to the length–tension relationship of muscle, an increase in the length of a resting (unstimulated) muscle will increase the force of contraction when the muscle is stimulated to contract. Therefore the preload force acts to augment the force of muscle contraction. In the intact heart, the stretch imposed on the cardiac muscle prior to the onset of muscle contraction is a function of the volume in the ventricles at the end of diastole. Therefore the end-diastolic volume of the ventricles is the preload force of the intact heart (3).

Preload and Systolic Performance The pressure-volume curves in Figure 1.1 show the influence of diastolic volume on the systolic performance of the heart. As the ventricle fills during diastole, there is an increase in both diastolic and systolic pressures. The increase in diastolic pressure is a reflection of the passive stretch imposed on the ventricle, while the difference between diastolic and systolic pressures is a reflection of the strength of ventricular contraction. Note that as diastolic volume increases, there is an increase in the difference between diastolic and systolic pressures, indicating that the strength of ventricular contraction is increasing. The importance of preload in augmenting cardiac contraction was discovered independently by Otto Frank (a German engineer) and Ernest Starling (a British physiologist), and their discovery is commonly referred to as the Frank-Starling relationship of the heart (3). This relationship can be stated as follows: In the normal heart, diastolic volume is the principal force that governs the strength of

ventricular contraction (3).

Clinical Monitoring In the clinical setting, the relationship between preload and systolic performance is monitored with ventricular function curves like the ones P.5 shown in Figure 1.2. End-diastolic pressure (EDP) is used as the clinical measure of preload because end-diastolic volume is not easily measured (the measurement of EDP is described in Chapter 10). The normal ventricular function curve has a steep ascent, indicating that changes in preload have a marked influence on systolic performance in the normal heart (i.e., the Frank-Starling relationship). When myocardial contractility is reduced, there is a decrease in the slope of the curve, resulting in an increase in end-diastolic pressure and a decrease in stroke volume. This is the hemodynamic pattern seen in patients with heart failure. Figure 1.1 Pressure-volume curves showing the influence of diastolic volume on the strength of ventricular contraction.

View Figure

Ventricular function curves are used frequently in the intensive care unit (ICU) to evaluate patients who are hemodynamically unstable. However, these curves can be misleading. The major problem is that conditions other than myocardial contractility can influence the slope of these curves. These conditions (i.e., ventricular compliance and ventricular afterload) are described next.

Preload and Ventricular Compliance The stretch imposed on cardiac muscle is determined not only by the volume of blood in the ventricles, but also by the tendency of the ventricular wall to distend or stretch in response to ventricular filling. P.6 The distensibility of the ventricles is referred to as compliance and can be derived using the following relationship between changes in end-diastolic pressure (EDP) and end-diastolic volume (EDV) (5):

Figure 1.2 Ventricular function curves used to describe the relationship between preload (end-diastolic pressure) and systolic performance (stroke volume).

View Figure

The pressure-volume curves in Figure 1.3 illustrate the influence of ventricular compliance on the relationship between ?EDP and ?EDV. As compliance decreases (i.e., as the ventricle becomes stiff), the slope of the curve decreases, resulting in a decrease in EDV at any given EDP. In this situation, the EDP will overestimate the actual preload (EDV). This illustrates how changes in ventricular compliance will influence the reliability of EDP as a reflection of preload. The following statements highlight the importance of ventricular compliance in the interpretation of the EDP measurement. 1. End-diastolic pressure is an accurate reflection of preload only when ventricular compliance is normal. 2. Changes in end-diastolic pressure accurately reflect changes in preload only when ventricular compliance is constant. Several conditions can produce a decrease in ventricular compliance. The most common are left ventricular hypertrophy and ischemic heart P.7 disease. Since these conditions are also commonplace in ICU patients, the reliability of the EDP measurement is a frequent concern.

Figure 1.3 Diastolic pressure-volume curves in the normal and noncompliant (stiff) ventricle.

View Figure

Diastolic Heart Failure As ventricular compliance begins to decrease (e.g., in the early stages of ventricular hypertrophy), the EDP rises, but the EDV remains unchanged. The increase in EDP reduces the pressure gradient for venous inflow into the heart, and this eventually leads to a decrease in EDV and a resultant decrease in cardiac output (via the Frank-Starling mechanism). This condition is depicted by the point on the lower graph in Figure 1.3, and is called diastolic heart failure (6). Systolic function (contractile strength) is preserved in this type of heart failure. Diastolic heart failure should be distinguished from conventional (systolic) heart failure because the management of the two conditions differs markedly. For example, since ventricular filling volumes are reduced in diastolic heart failure, diuretic therapy can be counterproductive. Unfortunately, it is not possible to distinguish between the two types of heart failure when the EDP is used as a measure of preload because the EDP is elevated in both conditions. The ventricular function curves in Figure 1.3 illustrate this problem. The point on the lower curve identifies a condition where EDP is elevated and stroke volume is reduced. This condition is often assumed to represent heart failure due to systolic dysfunction, but diastolic dysfunction would also produce the same changes. This inability to distinguish between systolic and diastolic heart failure is one of the major shortcomings of ventricular function curves. (See Chapter 14 for a more detailed discussion of systolic and diastolic heart failure.) P.8

Afterload When a weight is attached to one end of a contracting muscle, the force of muscle contraction must overcome the opposing force of the weight before the muscle begins to shorten. The weight in this situation represents a force called the afterload, which is defined as the load imposed on a muscle after the onset of muscle contraction. Unlike the preload force, which facilitates muscle contraction, the afterload force opposes muscle contraction (i.e., as the afterload increases, the muscle must develop more tension to move the load). In the intact heart, the afterload force is equivalent to the peak tension developed across the wall of the ventricles during systole (3). The determinants of ventricular wall tension (afterload) were derived from observations on soap bubbles made by the Marquis de Laplace in 1820. His observations are

expressed in the Law of Laplace, which states that the tension (T) in a thin-walled sphere is directly related to the chamber pressure (P) and radius (r) of the sphere: T = Pr. When the LaPlace relationship is applied to the heart, T represents the peak systolic transmural wall tension of the ventricle, P represents the transmural pressure across the ventricle at the end of systole, and r represents the chamber radius at the end of diastole ( 5). The forces that contribute to ventricular afterload can be identified using the components of the Laplace relationship, as shown in Figure 1.4. There are three major contributing forces: pleural pressure, arterial impedance, and end-diastolic volume (preload). Preload is a component of afterload because it is a volume load that must be moved by the ventricle during systole. Figure 1.4 The forces that contribute to ventricular afterload.

View Figure


Pleural Pressure Since afterload is a transmural force, it is determined in part by the pleural pressure on the outer surface of the heart. Negative pleural pressures will increase transmural pressure and increase ventricular afterload, while positive pleural pressures will have the opposite effect. Negative pressures surrounding the heart can impede ventricular emptying by opposing the inward displacement of the ventricular wall during systole (7,8). This effect is responsible for the transient decrease in systolic blood pressure (reflecting a decrease in cardiac stroke volume) that normally occurs during the inspiratory phase of spontaneous breathing. When the inspiratory drop in systolic pressure is greater than 15 mm Hg, the condition is called “pulsus paradoxus” (which is a misnomer, since the response is not paradoxical, but is an exaggeration of the normal response). Positive pleural pressures can promote ventricular emptying by facilitating the inward movement of the ventricular wall during systole (7,9). This effect is illustrated in Figure 1.5. The tracings in this figure show the effect of positive-pressure mechanical ventilation on the arterial blood pressure. When intrathoracic pressure rises during a positive-pressure breath, there is a transient rise in systolic blood pressure (reflecting an increase in the stroke volume output of the heart). This response indicates that positive intrathoracic pressure can provide P.10

cardiac support by “unloading” the left ventricle. Although this effect is probably of minor significance, positive-pressure mechanical ventilation has been proposed as a possible therapeutic modality in patients with cardiogenic shock (10). The hemodynamic effects of mechanical ventilation are discussed in more detail in Chapter 24. Figure 1.5 Respiratory variations in blood pressure during positive-pressure mechanical ventilation.

View Figure

Impedance The principal determinant of ventricular afterload is a hydraulic force known as impedance that opposes phasic changes in pressure and flow. This force is most prominent in the large arteries close to the heart, where it acts to oppose the pulsatile output of the ventricles. Aortic impedance is the major afterload force for the left ventricle, and pulmonary artery impedance serves the same role for the right ventricle. Impedance is influenced by two other forces: (a) a force that opposes the rate of change in flow, known as compliance, and (b) a force that opposes steady flow, called resistance. Arterial compliance is expressed primarily in the large, elastic arteries, where it plays a major role in determining vascular impedance. Arterial resistance is expressed primarily in the smaller peripheral arteries, where the flow is steady and nonpulsatile. Since resistance is a force that opposes nonpulsatile flow, while impedance opposes pulsatile flow, arterial resistance may play a minor role in the impedance to ventricular emptying. Arterial resistance can, however, influence pressure and flow events in the large, proximal arteries (where impedance is prominent) because it acts as a downstream resistance for these arteries. Vascular impedance and compliance are complex, dynamic forces that are not easily measured (12,13). Vascular resistance, however, can be calculated as described next.

Vascular Resistance The resistance (R) to flow in a hydraulic circuit is expressed by the relationship between the pressure gradient across the circuit (?P) and the rate of flow (Q) through the circuit:

Applying this relationship to the systemic and pulmonary circulations yields the following equations for systemic vascular resistance (SVR) and pulmonary vascular resistance


SAP is the mean systemic arterial pressure, RAP is the mean right atrial pressure, PAP is mean pulmonary artery pressure, LAP is the mean left atrial pressure, and CO is the cardiac output. The SAP is measured with an arterial catheter (see Chapter 8), and the rest of the measurements are obtained with a pulmonary artery catheter (see Chapter 9). P.11

Clinical Monitoring There are no accurate measures of ventricular afterload in the clinical setting. The SVR and PVR are used as clinical measures of afterload, but they are unreliable (14,15). There are two problems with the use of vascular resistance calculations as a reflection of ventricular afterload. First, arterial resistance may contribute little to ventricular afterload because it is a force that opposes nonpulsatile flow, while afterload (impedance) is a force that opposes pulsatile flow. Second, the SVR and PVR are measures of total vascular resistance (arterial and venous), which is even less likely to contribute to ventricular afterload than arterial resistance. These limitations have led to the recommendation that PVR and SVR be abandoned as clinical measures of afterload (15). Since afterload can influence the slope of ventricular function curves (see Figure 1.2), changes in the slope of these curves are used as indirect evidence of changes in afterload. However, other forces, such as ventricular compliance and myocardial contractility, can also influence the slope of ventricular function curves, so unless these other forces are held constant, a change in the slope of a ventricular function curve cannot be used as evidence of a change in afterload.

Contractility The contraction of striated muscle is attributed to interactions between contractile proteins arranged in parallel rows in the sarcomere. The number of bridges formed between adjacent rows of contractile elements determines the contractile state or contractility of the muscle fiber. The contractile state of a muscle is reflected by the force and velocity of muscle contraction when loading conditions (i.e., preload and afterload) are held constant (3). The standard measure of contractility is the acceleration rate of ventricular pressure (dP/dt) during isovolumic contraction (the time from the onset of systole to the opening of the aortic valve, when preload and afterload are constant). This can be measured during cardiac catheterization.

Clinical Monitoring There are no reliable measures of myocardial contractility in the clinical setting. The relationship between end-diastolic pressure and stroke volume (see Figure 1.2) is often used as a reflection of contractility; however, other conditions (i.e., ventricular compliance and afterload) can influence this relationship. There are echocardiography techniques for evaluating contractility (15,16), but these are very specialized and not used routinely.

Peripheral Blood Flow

As mentioned in the introduction to this chapter, there are over 60,000 miles of blood vessels in the human body! Even if this estimate is off by 10,000 or 20,000 miles, it still points to the incomprehensible vastness of the human circulatory system. The remainder of this chapter will describe the forces that govern flow through this vast network of blood vessels. P.12 A Note of Caution: The forces that govern peripheral blood flow are derived from observations on idealized hydraulic circuits where the flow is steady and laminar (streamlined), and the conducting tubes are rigid. These conditions bear little resemblance to the human circulatory system, where the flow is often pulsatile and turbulent, and the blood vessels are compressible and not rigid. Because of these differences, the description of blood flow that follows should be viewed as a very schematic representation of what really happens in the circulatory system.

Flow in Rigid Tubes Steady flow (Q) through a hollow, rigid tube is proportional to the pressure gradient along the length of the tube (?P), and the constant of proportionality is the hydraulic resistance to flow (R): The resistance to flow in small tubes was described independently by a German physiologist (G. Hagen) and a French physician (J. Poisseuille). They found that resistance to flow is a function of the inner radius of the tube (r), the length of the tube (L), and the viscosity of the fluid (m). Their observations are expressed in the following equation, known as the Hagen-Poisseuille equation (18): The final term in the equation is the reciprocal of resistance (1/R), so resistance can be described as

The Hagen-Poisseuille equation is illustrated in Figure 1.6. Note that flow varies according to the fourth power of the inner radius of the tube. This means that a two-fold increase in the radius of the tube will result in a sixteen-fold increase in flow: (2r)4 = 16r. The other components of resistance (i.e., tube length and fluid viscosity) exert a much smaller influence on flow. Since the Hagen-Poisseuille equation describes steady flow through rigid tubes, it may not accurately describe the behavior of the circulatory system (where flow is not steady and the tubes are not rigid). However, there are several useful applications of this equation. In Chapter 6, it will be used to describe flow through vascular catheters (see Figure 6.1). In Chapter 12, it will be used to describe the flow characteristics of different resuscitation fluids, and in Chapter 36, it will be used to describe the hemodynamic effects of anemia and blood transfusions.

Flow in Tubes of Varying Diameter As blood moves away from the heart and encounters vessels of decreasing diameter, the resistance to flow should increase and the flow should decrease. This is not possible because (according to the principle of

P.13 continuity) blood flow must be the same at all points along the circulatory system. This discrepancy can be resolved by considering the influence of tube narrowing on flow velocity. For a rigid tube of varying diameter, the velocity of flow (v) at any point along the tube is directly proportional to the bulk flow (Q), and inversely proportional to the cross-sectional area of the tube: v = Q/A (2). Rearranging terms (and using A = p 2) yields the following:

Figure 1.6 The forces that influence steady flow in rigid tubes. Q = flow rate, Pin = inlet pressure, Pout = outlet pressure, µ = viscosity, r = inner radius, L = length.

View Figure

This shows that bulk flow can remain unchanged when a tube narrows if there is an appropriate increase in the velocity of flow. This is how the nozzle on a garden hose works and is how blood flow remains constant as the blood vessels narrow.

Flow in Compressible Tubes Flow through compressible tubes (like blood vessels) is influenced by the external pressure surrounding the tube. This is illustrated in Figure 1.7, which shows a compressible tube running through a fluid reservoir. The height of the fluid in the reservoir can be adjusted to vary the external pressure on the tube. When there is no fluid in the reservoir and the external pressure is zero, the driving force for flow through the tube will be the pressure gradient between the two ends of the tube (P in - P out). When the reservoir fills and the external pressure exceeds the lowest pressure in the tube (P ext – P out), the tube will be compressed. In this situation, the driving force for flow is the pressure gradient between the inlet pressure and the external pressure (P in - Pext ). Therefore when a tube is compressed by external pressure, the driving force for flow is independent of the pressure gradient along the tube (20).

Figure 1.7 The influence of external pressure on flow through compressible tubes. Pin = inlet pressure, Pout = outlet pressure, Pext = external pressure.

View Figure


The Pulmonary Circulation Vascular compression has been demonstrated in the cerebral, pulmonary, and systemic circulations. It can be particularly prominent in the pulmonary circulation during positive-pressure mechanical ventilation, when alveolar pressure exceeds the hydrostatic pressure in the pulmonary capillaries (20). When this occurs, the driving force for flow through the lungs is no longer the pressure gradient from the main pulmonary arteries to the left atrium (PAP - LAP), but instead is the pressure difference between the pulmonary artery pressure and the alveolar pressure (PAP - Palv). This change in driving pressure not only contributes to a reduction in pulmonary blood flow, but it also affects the pulmonary vascular resistance (PVR) calculation as follows:

Vascular compression in the lungs is discussed again in Chapter 10 (the measurement of vascular pressures in the thorax) and in Chapter 24 (the hemodynamic effects of mechanical ventilation).

Blood Viscosity A solid will resist being deformed (changing shape), while a fluid will deform continuously (flow) but will resist changes in the rate of deformation (i.e., changes in flow rate) (21). The resistance of a fluid to P.15 changes in flow rate is a property known as viscosity (21,22,23). Viscosity has also been referred to as the “gooiness” of a fluid (21). When the viscosity of a fluid increases, a greater force must be applied to the fluid to initiate a change in flow rate. The influence of viscosity on flow rate is apparent to anyone who has poured molasses (high viscosity) and water (low viscosity) from a container.

Hematocrit The viscosity of whole blood is almost entirely due to cross-linking of circulating erythrocytes by plasma fibrinogen (22,23). The principal determinant of whole blood

viscosity is the concentration of circulating erythrocytes (the hematocrit). The influence of hematocrit on blood viscosity is shown in Table 1.2. Note that blood viscosity can be expressed in absolute or relative terms (relative to water). In the absence of blood cells (zero hematocrit), the viscosity of blood (plasma) is only slightly higher than that of water. This is not surprising, since plasma is 92% water. At a normal hematocrit (45%), blood viscosity is three times the viscosity of plasma. Thus plasma flows much more easily than whole blood, and anemic blood flows much more easily than normal blood. The influence of hematocrit on blood viscosity is the single most important factor that determines the hemodynamic effects of anemia and blood transfusions (see later).

Shear Thinning The viscosity of some fluids varies inversely with a change in flow velocity (21,23). Blood is one of these fluids. (Another is ketchup, which is thick and difficult to get out of the bottle, but once it starts to flow, it thins out and flows more easily.) Since the velocity of blood flow increases as the blood vessels narrow, the viscosity of blood will also decrease as the P.16 blood moves into the small blood vessels in the periphery. The decrease in viscosity occurs because the velocity of plasma increases more than the velocity of erythrocytes, so the relative plasma volume increases in small blood vessels. This process is called shear thinning (shear is a tangential force that influences flow rate), and it facilitates flow through small vessels. It becomes evident in blood vessels with diameters less than 0.3 mm (24). TABLE 1.2 Blood Viscosity as a Function of Hematocrit


Viscosity* Relative

























* Absolute viscosity expressed in centipoise (cP). Data from Documenta Geigy Scientific Tables. 7th ed. Basel: Documenta Geigy, 1966:557–558.

Hemodynamic Effects The Hagen-Poisseuille equation indicates that blood flow is inversely related to blood viscosity, and further that blood flow will change in proportion to a change in viscosity (i.e., if blood viscosity is doubled, blood flow will be halved) (22). The effect of changes in blood viscosity on blood flow is shown in Figure 1.8. In this case, changes in hematocrit are used to represent changes in blood viscosity. The data in this graph is from a patient with polycythemia who was treated with a combination of phlebotomy and fluid infusion (isovolemic hemodilution) to achieve a therapeutic reduction in hematocrit and blood viscosity. The progressive decrease in hematocrit is associated with a steady rise in cardiac output, and the change in cardiac output is far greater than the change in hematocrit. The disproportionate increase in cardiac output is more than expected from the Hagen-Poisseuille equation and may be due in part P.17 to the fact that blood viscosity varies inversely with flow rate. That is, as viscosity decreases and flow rate increases, the increase in flow rate will cause a further reduction

in viscosity, which will then lead to a further increase in flow rate, and so on. This process would then magnify the influence of blood viscosity on blood flow. Whether or not this is the case, the graph in Figure 1.8 demonstrates that changes in hematocrit have a profound influence on circulatory blood flow. This topic is presented in more detail in Chapter 36. Figure 1.8 The influence of progressive hemodilution on cardiac output in a patient with polycythemia. CO = cardiac output. (From LeVeen HH, Ip M, Ahmed N, et al. Lowering blood viscosity to overcome vascular resistance. Surg Gynecol Obstet 1980;150:139.Bibliographic Links)

View Figure

Clinical Monitoring Viscosity can be measured with an instrument called (what else?) a viscometer. This device has two parallel plates: one fixed and one that can move over the surface of the fixed plate. A fluid sample is placed between the two plates, and a force is applied to move the moveable plate. The force needed to move the plate is proportional to the viscosity of the fluid between the plates. Viscosity is expressed as force per area (surface area of the plates). The units of measurement are the “poise” (or dyne · sec/cm 2) in the CGS system, and the “Pascal · second” (Pa · s) in the SI system. (A poise is one/tenth of a Pascal · second.) Viscosity is also expressed as the ratio of the test sample viscosity to the viscosity of water. This “relative viscosity” is easier to interpret. Viscosity is rarely measured in the clinical setting. The main reason for this is the consensus view that in vitro viscosity measurements are unreliable because they do not take into account conditions in the circulatory system (like shear thinning) that influence viscosity (21,22,23,24). Monitoring changes in viscosity may be more useful than single measurements. For example, serial changes in blood viscosity could be used to monitor the effects of aggressive diuretic therapy (e.g., a rise in viscosity to abnormally high levels might trigger a reduction in diuretic dosage). The value of blood viscosity measurements is underappreciated at the present time.

References General Texts

Berne R, Levy M. Cardiovascular physiology, 8th ed. St. Louis: Mosby, 2001. Guyton AC, Jones CE, Coleman TG. Circulatory physiology: cardiac output and its regulation, 2nd ed. Philadelphia: WB Saunders, 1973. Nichols WW, O'Rourke M. McDonald's blood flow in arteries, 3rd ed. Baltimore: Williams & Wilkins, 1990. Vogel S. Vital circuits. New York: Oxford University Press, 1992. Warltier DC. Ventricular function. Baltimore: Williams & Wilkins, 1995.

Cardiac Output 1. Vogel S. Vital circuits. New York: Oxford University Press, 1992:1–17.

2. Vogel S. Life in moving fluids. Princeton: Princeton University Press, 1981: 25–28. P.18 3. Opie LH. Mechanisms of cardiac contraction and relaxation. In: Braunwald E, Zipes DP, Libby P, eds. Heart disease: a textbook of cardiovascular medicine, 6th ed. Philadelphia: WB Saunders, 2001:443–478.

4. Parmley WM, Talbot L. The heart as a pump. In: Berne RM, ed. Handbook of physiology: the cardiovascular system. Bethesda: American Physiological Society, 1979:429–460. 5. Gilbert JC, Glantz SA. Determinants of left ventricular filling and of the diastolic pressure-volume relation. Circ Res 1989;64:827–852. Ovid Full TextBibliographic Links 6. Zile M, Baicu C, Gaasch W. Diastolic heart failure: abnormalities in active relaxation and passive stiffness of the left ventricle. N Engl J Med 2004;350: 1953–1959. Ovid Full TextBibliographic Links 7. Pinsky MR. Cardiopulmonary interactions: the effects of negative and positive changes in pleural pressures on cardiac output. In: Dantzger DR, ed. Cardiopulmonary critical care, 2nd ed. Philadelphia: WB Saunders, 1991: 87–120.

8. Hausnecht N, Brin K, Weisfeldt M, et al. Effects of left ventricular loading by negative intrathoracic pressure in dogs. Circ Res 1988;62:620–631. Ovid Full TextBibliographic Links 9. Magder S. Clinical usefulness of respiratory variations in blood pressure. Am J Respir Crit Care Med 2004;169:151–155. Full TextBibliographic Links 10. Peters J. Mechanical ventilation with PEEP: a unique therapy for failing hearts. Intens Care Med 1999;25:778–780. Bibliographic Links 11. Nichols WW, O'Rourke MF. Input impedance as ventricular load. In: McDonald's blood flow in arteries, 3rd ed. Philadelphia: Lea & Febiger, 1990: 330–342. 12. Finkelstein SM, Collins R. Vascular impedance measurement. Progr Cardio-vasc Dis 1982;24:401–418. Full TextBibliographic Links 13. Laskey WK, Parker G, Ferrari VA, et al. Estimation of total systemic arterial compliance in humans. J Appl Physiol 1990;69:112–119. Bibliographic Links 14. Lang RM, Borrow KM, Neumann A, et al. Systemic vascular resistance: an unreliable index of left ventricular afterload. Circulation 1986;74:1114–1123. Ovid Full TextBibliographic Links 15. Pinsky MR. Hemodynamic monitoring in the intensive care unit. Clin Chest Med 2003;24:549–560. Bibliographic Links 16. Bargiggia GS, Bertucci C, Recusani F, et al. A new method for estimating left ventricular dP/dt by continuous wave Doppler echocardiography: validation studies at cardiac catheterization. Circulation 1989;80:1287–1292. Ovid Full TextBibliographic Links 17. Broka S, Dubois P, Jamart J, et al. Effects of acute decrease in afterload on accuracy of Doppler-derived left ventricular rate of pressure rise measurement in anesthetized patients. J Am Soc Echocardiogr 2001;14:1161–1165.Bibliographic Links

Peripheral Blood Flow

18. Chien S, Usami S, Skalak R. Blood flow in small tubes. In: Renkin EM, Michel CC, eds. Handbook of physiology, Section 2: the cardiovascular system. Vol IV: The microcirculation. Bethesda: American Physiological Society, 1984:217–249.

19. Little RC, Little WC. Physiology of the heart and circulation, 4th ed. Chicago: Year Book Publishers, 1989:219–236. 20. Gorback MS. Problems associated with the determination of pulmonary vascular resistance. J Clin Monit 1990;6:118–127. Bibliographic Links P.19 21. Vogel S. Life in moving fluids. Princeton: Princeton University Press, 1981: 11–24.

22. Merrill EW. Rheology of blood. Physiol Rev 1969;49:863–888. Bibliographic Links 23. Lowe GOD. Blood rheology in vitro and in vivo. Baillieres Clin Hematol 1987;1:597.

24. Berne RM, Levy MN. Cardiovascular physiology, 8th ed. Philadelphia: Mosby, 1992:127–133.

Chapter 2 Oxygen and Carbon Dioxide Transport Respiration is thus a process of combustion, in truth very slow, but otherwise exactly like that of charcoal. --Antoine Lavoisier The business of aerobic metabolism is the combustion of nutrient fuels to release energy. This process consumes oxygen and generates carbon dioxide. The business of the circulatory system is to deliver the oxygen and nutrient fuels to the tissues of the body, and then to remove the carbon dioxide that is generated. The dual role of the circulatory system in transporting both oxygen and carbon dioxide is referred to as the respiratory function of blood. The business of this chapter is to describe how this respiratory function is carried out.

Oxygen Transport The transport of oxygen from the lungs to metabolizing tissues can be described by using four clinical parameters: (a) the concentration of oxygen in blood, (b) the delivery rate of oxygen in arterial blood, (c) the rate of oxygen uptake from capillary blood into the tissues, and (d) the fraction of oxygen in capillary blood that is taken up into the tissues. These four oxygen transport parameters are shown in Table 2.1, along with the equations used to derive each parameter. Thorough knowledge of these parameters is essential for the management of critically ill patients.

O2 Content in Blood Oxygen does not dissolve readily in water (1) and, since plasma is 93% water, a specialized oxygen-binding molecule (hemoglobin) is needed to P.22 facilitate the oxygenation of blood. The concentration of oxygen (O2) in blood, also called the O2 content, is the summed contribution of O2 that is bound to hemoglobin and O 2 that is dissolved in plasma. TABLE 2.1 Oxygen and Carbon Dioxide Transport Parameters




Arterial O2 content

CaO 2

1.34 × Hb × SaO2

Venous O 2 content


1.34 × Hb × SvO 2

O2 Delivery


Q × CaO2

O2 Uptake


Q × (CaO2 – CvO2)

O2 Extraction ratio



CO2 Elimination


Q × (CvCO2 - CaCO 2)

Respiratory quotient



Abbreviations: Hb = hemoglobin concentration in blood; SaO 2 and SvO 2 = oxygen saturation of hemoglobin (ratio of oxygenated hemoglobin to total hemoglobin) in arterial and mixed venous blood, respectively; Q = cardiac output; CaCO2 = CO2 content in arterial blood; CvCO 2 = CO2 content in mixed venous blood.

Hemoglobin-Bound O2 The concentration of hemoglobin-bound O 2 (HbO 2) is determined by the variables in Equation 2.1 (2). Hb is the hemoglobin concentration in blood (usually expressed in grams per deciliter, which is grams per 100 mL); 1.34 is the oxygen-binding capacity of hemoglobin (expressed in mL O2 per gram of Hb); and SO2 is the ratio of oxygenated hemoglobin to total hemoglobin in blood (SO2 = HbO2/total Hb), also called the O2 saturation of hemoglobin. The HbO2 is expressed in the same units as the Hb concentration (g/dL). Equation 2.1 predicts that, when hemoglobin is fully saturated with O 2 (i.e., when the SO2 = 1), each gram of hemoglobin will bind 1.34 mL oxygen. One gram of hemoglobin normally binds 1.39 mL oxygen, but a small fraction (3% to 5%) of circulating hemoglobin is present as methemoglobin and carboxyhemoglobin and, since these forms of Hb have

a reduced O2-binding capacity, the lower value of 1.34 mL/g is considered more representative of the O2-binding capacity of the total hemoglobin pool (3).

Dissolved O2 The concentration of dissolved oxygen in plasma is determined by the solubility of oxygen in water (plasma) and the partial pressure of oxygen (PO 2) in blood. The solubility of O2 in water is temperature-dependent (solubility increases slightly as temperature decreases). At normal body temperature (37 °C), 0.03 mL of O2 will dissolve in one liter of water when the Po2 is 1 mm Hg (4). This is expressed as a solubility coefficient of 0.03 mL/L/mm Hg (or 0.003 mL/100 mL/mm Hg). The concentration of P.23 dissolved O2 (in mL/dL) (at normal body temperature) is then described by Equation 2.2.

TABLE 2.2 Normal Levels of Oxygen in Arterial and Venous Blood*


Arterial Blood

Venous Blood


90 mm Hg

40 mm Hg

O2 Saturation of Hb



Hb-bound O 2

197 mL/L

147 mL/L

Dissolved O2

2.7 mL/L

1.2 mL/L

Total O2 content

200 mL/L

148 mL/L

Blood volume †

1.25 L

3.75 L

Volume of O 2

250 mL

555 mL

*Values shown are for a body temperature of 37°C and a hemoglobin concentration of 15 g/dL (150 g/L) in blood. †Volume estimates are based on a total blood volume (TBV) of = L, arterial blood volume of 0.25 × TBV, and venous blood volume of 0.75 3 TBV. Abbreviations: Hb, hemoglobin 5 PO2, partial pressure of O2.

This equation reveals the limited solubility of oxygen in plasma. For example, if the Po 2 is 100 mm Hg, one liter of blood will contain only 3 mL of dissolved o2.

Arterial O2 Content (Cao2) The concentration of O2 in arterial blood (Cao 2) can be defined by combining Equations 2.1 and 2.2, by using the So2 and Po2 of arterial blood (Sao 2 and Pao2). The normal concentrations of bound, dissolved, and total O2 in arterial blood are shown in Table 2.2. There are approximately 200 mL oxygen in each liter of arterial blood, and only 1.5% (3 mL) is dissolved in the plasma. The oxygen consumption of an average-sized adult at rest is 250 mL/min, which means that if we were forced to rely solely on the dissolved O2 in plasma, a cardiac output of 89 L/min would be necessary to sustain aerobic metabolism. This emphasizes the importance of hemoglobin in the transport of oxygen.

Venous O2 Content (Cvo2) The concentration of O2 in venous blood (Cvo2) can be calculated in the same fashion as the Cao2, using the O 2 saturation and Po2 in venous blood (Svo2 and Pvo2). P.24 The Svo2 and Pvo2 are best measured in a pooled or “mixed venous” blood sample taken from the pulmonary artery (using a pulmonary artery catheter, as described in Chapter 9). As shown in Table 2.2, the normal Svo2 is 73% (0.73), the normal Pvo 2 is 40 mm Hg, and the normal Cvo2 is approximately 15 mL/dL (150 mL/L).

Simplified O2 Content Equation The concentration of dissolved O 2 in plasma is so small that it is usually eliminated from the O2 content equation. The O2 content of blood is then considered equivalent to the Hb-bound O 2 fraction (see Equation 2.1).

Anemia versus Hypoxemia Clinicians often use the arterial Po2 (Pao2) as an indication of how much oxygen is in the blood. However, as indicated in Equation 2.5, the hemoglobin concentration is the principal determinant of the oxygen content of blood. The comparative influence of hemoglobin and Pao2 on the oxygen level in blood is shown in Figure 2.1. The graph in this figure shows the effect of proportional changes in hemoglobin concentration and Pao2 on the oxygen content of arterial blood. A 50% reduction in hemoglobin (from 15 to 7.5 g/dL) is accompanied by an equivalent 50% reduction in Cao 2 (from 200 to 101 mL/L),

while a similar 50% reduction in the PaO 2 (from 90 to 45 mm Hg) results in only an 18% decrease in Cao2 (from 200 to 163 mL/L). This graph shows that anemia has a much more profound effect on blood oxygenation than hypoxemia. It should also serve as a reminder to avoid using the Pao2 to assess arterial oxygenation. The Pao 2 should be used to evaluate the efficiency of gas exchange in the lungs (see Chapter 19).

The Paucity of O2 in Blood The total volume of O 2 in circulating blood can be calculated as the product of the blood volume and the O2 concentration in blood. An estimate of the volume of O 2 in arterial and venous blood is shown in Table 2.2. The combined volume of O 2 in arterial and venous blood is a meager 805 mL. To appreciate what a limited volume this represents, consider that the whole-body O2 consumption of an average-sized adult at rest is about 250 mL/min. This means that the total volume of O2 in blood is enough to sustain aerobic metabolism for only 3 to 4 minutes. Thus if a patient stops breathing, you have only a precious few minutes to begin assisted breathing maneuvers before the oxygen stores in the blood are completely exhausted. The limited quantity of O2 in blood can also be demonstrated by considering the oxidative metabolism of glucose, which is described by the formula: C 6H12O6 + 6O2 ? 6CO 2 + 6H2O. This formula indicates that complete oxidation of one mole of glucose utilizes 6 moles of oxygen. To determine if the O2 in blood is enough to metabolize the glucose in blood, it is necessary to express the amount of glucose and oxygen in blood in P.25 millimoles (mmol). (The values shown here are based on a blood glucose level of 90 mg/dL or 90/180 = 0.5 mmol/dL, a blood volume of 5 liters, and a total blood O2 of 805 mL or 805/22.4 = 36.3 mmol): Total glucose in blood………………… 25 mmol Total O2 in blood……………………… 36.3 mmol O2 need of glucose metabolism…………… 150 mmol

Figure 2.1 Graph showing the effects of equivalent (50%) reductions in hemoglobin concentration (Hb) and arterial Po2 (Pao2) on the oxygen concentration in arterial blood (Cao2).

View Figure

This shows that the O 2 in blood is only about 20% to 25% of the amount needed for the complete oxidative metabolism of the glucose in blood.

Why so Little O2? The obvious question is why an organism that requires oxygen for survival is designed to carry on metabolism in an oxygen-limited environment? The answer may be related to the toxic potential of oxygen. Oxygen is well known for its ability to produce lethal cell injury via the production of toxic metabolites (superoxide radical, hydrogen peroxide, P.26 and the hydroxyl radical), so limiting the oxygen concentration in the vicinity of cells may be a mechanism for protecting cells from oxygen-induced cell injury. The role of oxygen-induced injury (oxidant injury) in clinical disease is a very exciting and active area of study, and the bibliography at the end of this chapter includes a textbook (Free Radicals in Biology and Medicine) that is the best single source of information on this subject.

The Abundance of Hemoglobin In contrast to the small volume of oxygen in blood, the total mass of circulating hemoglobin seems excessively large. If the normal serum Hb is 15 g/dL (150 g/L) and the normal blood volume is 5 liters (70 mL/kg), the total mass of circulating hemoglobin is 750 grams (0.75 kg) or 1.65 lbs. To demonstrate the enormous size of the blood hemoglobin pool, the illustration in Figure 2.2 compares the mass of hemoglobin to the normal P.27 weight of the heart. The heart weighs only 300 grams, so the pool of circulating hemoglobin is 2.5 times heavier than the heart! This means that every 60 seconds, the heart must move a mass that is more than twice its own weight through the circulatory system.

Figure 2.2 A balance scale demonstrating the excess weight of circulating hemoglobin when matched with the normal weight of the heart. The numbers on the small weights indicate the weight of each in grams.

View Figure

Is all this hemoglobin necessary? As shown later, when the extraction of oxygen from the systemic capillaries is maximal, 40% to 50% of the hemoglobin in venous blood remains fully saturated with oxygen. This means that almost half of the circulating hemoglobin is not used to support aerobic metabolism. What is the excess hemoglobin doing? Transporting carbon dioxide, as described later in the chapter.

Oxygen Delivery (DO2) The oxygen that enters the bloodstream in the lungs is carried to the vital organs by the cardiac output. The rate at which this occurs is called the oxygen delivery (Do2). The Do 2 describes the volume of oxygen (in milliliters) that reaches the systemic capillaries each minute. It is equivalent to the product of the O2 content in arterial blood (Cao2) in mL/L and the cardiac output (Q) in L/min (2,5,6,7). (The multiplier of 10 is used to convert the Cao 2 from mL/dL to mL/L, so the DO2 can be expressed in mL/min.) If the Cao 2 is broken down into its components (1.34 3 Hb 3 SaO2), Equation 2.6 can be rewritten as When a pulmonary artery catheter is used to measure cardiac output (see Chapter 9), the Do2 can be calculated by using Equation 2.7. The normal Do 2 in adults at rest is 900–1,100 mL/min, or 500–600 mL/min/m 2 when adjusted for body size (see Table 2.3). TABLE 2.3 Normal Ranges for Oxygen and Carbon Dioxide Transport Parameters


Absolute Range

Size-Adjusted Range*

Cardiac output

5–6 L/min

2.4–4.0 L/min/m 2

O2 Delivery

900–1,100 mL/min

520–600 mL/min/m 2

O2 Uptake

200–270 mL/min

110–160 mL/min/m 2

O2 Extraction ratio


CO2 Elimination

160–220 mL/min

Respiratory quotient


90–130 mL/min/m2

*Size-adjusted values are the absolute values divided by the patient's body surface area in square meters (m 2).


Oxygen Uptake (VO2) When blood reaches the systemic capillaries, oxygen dissociates from hemoglobin and moves into the tissues. The rate at which this occurs is called the oxygen uptake (Vo2). The Vo2 describes the volume of oxygen (in mL) that leaves the capillary blood and moves into the tissues each minute. Since oxygen is not stored in tissues, the Vo2 is also a measure of the oxygen consumption of the tissues. The Vo 2 (in mL/min) can be calculated as the product of the cardiac output (Q) and the arteriovenous oxygen content difference (Cao 2 - Cvo 2). (The multiplier of 10 is included for the same reason as explained for the DO 2.) This method of deriving VO 2 is called the reverse Fick method because Equation 2.8 is a variation of the Fick equation (where cardiac output is the derived variable: Q = VO2/CaO2 - CvO2) (8). Since the CaO2 and CvO2 share a common term (1.34 × Hb × 10), Equation 2.8 can be restated as

This equation expresses VO 2 using variables that can be measured in clinical practice. The determinants of VO2 in this equation are illustrated in Figure 2.3. The normal range for VO2 in healthy adults at rest is 200–300 mL/min, or 110–160 mL/min/m2 when adjusted for body size (see Table 2.3).

View Figure

Figure 2.3 A schematic representation of the factors that determine the rate of oxygen uptake (VO 2) from the microcirculation. SaO2 and SvO 2 = Oxygen saturation of hemoglobin in arterial and venous blood, respectively; PO2 = partial pressure of oxygen; Hb = a hemoglobin molecule.


Fick vs Whole-Body VO2 The VO2 in the modified Fick equation is not equivalent to the whole-body VO2 because it does not include the O2 consumption of the lungs (8,9,10). Normally, the VO2 of the lungs represents less than 5% of the whole-body VO 2 (9), but it can make up 20% of the whole-body VO 2 in patients with inflammatory conditions in the lungs (which are common in ICU patients) (10). This discrepancy can be important when VO 2 is used as an end-point of hemodynamic management (see Chapter 11) because an underestimate of whole-body VO 2 could lead to overaggressive management to augment the VO2. Direct measurement of the VO2 (described next) is a more accurate representation of the whole-body VO 2.

Direct Measurement of VO2 The whole-body VO 2 can be measured directly by monitoring the rate of oxygen disappearance from the lungs. This can be accomplished with a specialized instrument equipped with an oxygen gas analyzer that is connected to the proximal airway (usually in intubated patients) to measure the O 2 concentration in inhaled and exhaled gas. The device records and displays the VO 2 as the product of minute ventilation (V E) and the fractional concentration of oxygen in inhaled and exhaled gas (F iO2 and FeO2). The direct measurement of VO2 is more accurate than the calculated (Fick) VO 2 because it is a closer approximation to the whole-body VO 2. It has several other advantages over the Fick VO2, and these are described in Chapter 11. The major shortcoming of the direct VO2 measurement is the lack of availability of monitoring equipment in many ICUs, and

the need for trained personnel to operate the equipment.

Oxygen-Extraction Ratio (O2ER) The fraction of the oxygen delivered to the capillaries that is taken up into the tissues is an index of the efficiency of oxygen transport. This is monitored with a parameter called the oxygen extraction ratio (O2ER), which is the ratio of O2 uptake to O2 delivery.

This ratio can be multiplied by 100 and expressed as a percentage. Since the VO 2 and DO2 share common terms (Q × 1.34 × Hb × 10), Equation 2.11 can be reduced to an equation with only two measured variables: When the SaO 2 is close to 1.0 (which is usually the case), the O 2ER is roughly equivalent to the (SaO2 - SvO2) difference: O2ER ˜ SaO2 - SvO2. P.30 The O2ER is normally about 0.25 (range = 0.2–0.3), as shown in Table 2.3. This means that only 25% of the oxygen delivered to the systemic capillaries is taken up into the tissues. Although O2 extraction is normally low, it is adjustable and can be increased when oxygen delivery is impaired. The adjustability of O2 extraction is an important factor in the control of tissue oxygenation, as described next.

Control of Oxygen Uptake The oxygen transport system operates to maintain a constant flow of oxygen into the tissues (a constant VO 2) in the face of changes in oxygen supply (varying DO 2). This behavior is made possible by the ability of O 2 extraction to adjust to changes in O 2 delivery (11). The control system for VO2 can be described by rearranging the O 2 extraction equation (Equation 2.11) to make VO2 the dependent variable:

This equation shows that the VO 2 will remain constant if changes in O2 delivery are accompanied by equivalent and reciprocal changes in O 2 extraction. However, if the O2 extraction remains fixed, changes in DO 2 will be accompanied by equivalent changes in VO2. The ability of O2 extraction to adjust to changes in DO2 therefore determines the ability to maintain a constant VO 2.

The DO2–VO2 Relationship The normal relationship between O 2 delivery and O2 uptake is shown in the graph in Figure 2.4 (11). As O2 delivery (DO2) begins to decrease below normal (as indicated by the arrowhead on the graph), the O 2 uptake (VO2) initially remains constant, indicating

that the O2 extraction (O2ER) is increasing as the DO 2 decreases. Further decreases in DO2 eventually leads to a point where the VO2 begins to decrease. The transition from a constant to a varying VO 2 occurs when the O2 extraction increases to a maximum level of 50% to 60% (O2ER = 0.5 to 0.6). Once the O2ER is maximal, further decreases in DO 2 will result in equivalent decreases in VO2 because the O2ER is fixed and cannot increase further. When this occurs, the VO2 is referred to as being supply-dependent, and the rate of aerobic metabolism is limited by the supply of oxygen. This condition is known as dysoxia (12). As aerobic metabolism (VO 2) begins to decrease, the oxidative production of high energy phosphates (ATP) begins to decline, resulting in impaired cell function and eventual cell death. The clinical expression of this process is clinical shock and progressive multiorgan failure (13).

The Critical DO2 The DO2 at which the VO2 becomes supply-dependent is called the critical oxygen delivery (critical DO2). It is the lowest DO2 that is capable of P.31 fully supporting aerobic metabolism and is identified by the bend in the DO 2–VO2 curve (see Fig. 2.4). Despite the ability to identify the anaerobic threshold, the critical DO2 has limited clinical value. First, the critical DO2 has varied widely in studies of critically ill patients (11,13,14), and it is not possible to predict the critical DO2 in any individual patient in the ICU. Second, the DO2–VO2 curve can be curvilinear (i.e., without a single transition point from constant to changing VO2) (15), and in these cases, it is not possible to identify a critical DO2. Figure 2.4 Graph showing the normal relationship between O 2 delivery (DO 2) and O2 uptake (VO2) when O 2 delivery is decreased progressively, as indicated by the arrowheads.

View Figure

The DO2:VO2 ratio may be a more useful parameter than the critical DO 2 for identifying (and avoiding) the anaerobic threshold. Maintaining a DO 2:VO2 ratio of 4:1 or higher has been recommended as a management strategy to avoid the anaerobic threshold in critically ill patients (7).

Carbon Dioxide Transport

Carbon dioxide (CO 2) is the major end-product of oxidative metabolism, and because it readily hydrates to form carbonic acid, it can be a source of significant acidosis if allowed to accumulate. The importance of eliminating CO2 from the body is apparent in the behavior of the ventilatory control system, which operates to maintain a constant PO 2 in arterial blood (PaCO2). An increase in PaCO2 of 5 mm Hg can result in a twofold increase in minute ventilation. To produce an equivalent increment in ventilation, the arterial PO2 must drop to 55 mm Hg (16). The tendency P.32 for the ventilatory control system to pay attention to hypercapnia and ignore hypoxemia is intriguing because it suggests that the ventilatory system is more concerned with eliminating metabolic waste (CO 2) than promoting aerobic metabolism (by supplying oxygen).

The Hydration of CO2 The total body CO2 in adults is reported at 130 liters (17), which doesn't seem possible in light of the fact that the total body water of an adult averages only 40 to 45 liters. This dilemma can be explained by the tendency for CO2 to enter into a chemical reaction with water and produce carbonic acid. The hydration of CO 2 and its transformation to carbonic acid is a continuous process, and this creates a perpetual gradient that drives CO 2 into solution. Since the CO 2 is continuously disappearing, the total volume of CO 2 in the solution could exceed the volume of the solution. If you have ever opened a bottle of warm champagne, you have witnessed how much CO 2 can be dissolved in a solution.

CO2 Transport Scheme The transport of CO 2 is a complex process that is shown in Figure 2.5. The centerpiece of CO2 transport is the reaction of CO2 with water. The first stage of this reaction involves the formation of carbonic acid. This is normally a slow reaction and takes about 40 seconds to complete (18). The reaction speeds up considerably in the presence of the enzyme carbonic anhydrase and takes less than 10 milliseconds (msec) to complete (18). Carbonic anhydrase is confined to the red cell and is not present in P.33 plasma. Thus CO2 is rapidly hydrated only in the red blood cell, and this creates a pressure gradient that drives CO 2 into the cell.

Figure 2.5 The chemical reactions involved in CO 2 transport. Values in parentheses indicate the amount of each component normally present in 1 L of venous blood. The double arrows indicate favored pathways. View Figure

Carbonic acid dissociates instantaneously to produce hydrogen and bicarbonate ions. A large fraction of the bicarbonate generated in the red cell is pumped back into the plasma in exchange for chloride. The hydrogen ion generated in the red cell is buffered by the hemoglobin. In this way, the CO2 that enters the red cell is dismantled and the parts stored (hemoglobin) or discarded (bicarbonate) to create room for more CO 2 to enter the red cell. These processes create a sink to accommodate large volumes of CO 2 in the red cell. A small fraction of CO2 in the red cell reacts with free amino groups on hemoglobin to produce carbamic acid, which dissociates to form carbamino residues (HbNHCOO) and hydrogen ions. This reaction provides another opportunity for hemoglobin to act as a buffer.

CO2 Content of Blood The different measures of CO2 in blood are listed in Table 2.4. Like oxygen, CO 2 is present in a dissolved form, and the concentration of dissolved CO 2 is determined as the product of the PCO 2 and the solubility coefficient for CO2 in water (i.e., 0.69 mL/L/mm Hg at 37°C) (19). The dissolved CO 2 content in arterial and venous blood is shown in Table 2.4 (20). Like oxygen, the dissolved CO2 is only a small fraction of the total CO 2 content of blood. The total content of CO2 in blood is the summed contribution of several components, including the dissolved CO 2 and bicarbonate concentrations in plasma and erythrocytes, and the carbamino CO 2 content in erythrocytes. The normal values for each of these components in venous blood are shown in Figure 2.5. If these values are summed, the total CO 2 content is 23 mEq/L, with 17 mEq/L in plasma and 6 mEq/L in the red cell. The preponderance of CO2 in plasma is deceiving because most of the plasma component is in the form of bicarbonate that has been expelled from the red blood cell. TABLE 2.4 Normal Levels of CO2 in Arterial and Venous Blood*


Arterial Blood

Venous Blood


40 mm Hg

45 mm Hg

Dissolved CO2

27 mL/L

29 mL/L

Total CO2 content

490 mL/L

530 mL/L

Blood volume †

1.25 L

3.75 L

Volume of CO2

613 mL

1,988 mL

*Values shown are for a body temperature of 37°C. †Volume estimates are based on a total blood volume (TBV) of 5 L, arterial blood volume of 0.25 × TBV, and venous blood volume of 0.75 3 TBV. Abbreviations: PCO2 = partial pressure of CO2.

TABLE 2.5 Buffering Capacity of Blood Proteins


Plasma Proteins

Inherent buffer capacity

0.18 mEq H+ /g

0.11 mEq H+ /g

Concentration in blood

150 g/L

38.5 g/L

Total buffer capacity

27.5 mEq H+ /L

4.2 mEq H+ /L

P.34 Because CO 2 readily dissociates into ions (hydrogen and bicarbonate), the concentration of CO2 is often expressed in ion equivalents (mEq/L), as in Figure 2.5. Conversion to units of volume (mL/L or mL/dL) is possible because one mole of CO 2 will occupy a

volume of 22.3 liters. Therefore: CO2 (mL/L) = CO2 (mEq/L × 22.3) Table 2.4 includes the CO2 content of blood expressed in volume units (20). Note that the total volume of CO2 in blood (about 2.6 liters) is more than 3 times the volume of O2 in blood (805 mL).

Hemoglobin As a Buffer Figure 2.5 shows that hemoglobin plays a central role in CO 2 transport by acting as a buffer for the hydrogen ions generated by the hydration of CO2 in the red blood cell. The buffering capacity of hemoglobin is shown in Table 2.5 (21). Note that the total buffering capacity of hemoglobin is six times greater than the combined buffering capacity of all the plasma proteins. The buffering actions of hemoglobin are attributed to imidazole groups that are found on the 38 histidine residues in the molecule. These imidazole groups have a dissociation constant with a pK of 7.0, so they will act as effective buffers in the pH range from 6 to 8 (buffers are effective within one pH unit on either side of the pK) (20). In contrast, the carbonic acid–bicarbonate buffer system has a pK of 6.1, so this buffer system will be effective in the pH range from 5.1 to 7.1. Comparing the buffer ranges of hemoglobin and bicarbonate shows that hemoglobin is a more effective buffer than bicarbonate in the pH range encountered clinically (pH 7 to 8)! This aspect of hemoglobin function deserves more attention.

Why the Excess Hemoglobin? As described earlier, the mass of hemoglobin in blood is far greater than needed to transport oxygen, and considering the role played by hemoglobin in CO 2 transport, it is likely that the excess hemoglobin is needed for CO 2 transport. Considering the large volume of CO2 in blood (see Table 2.4), it is easier to understand why there is so much hemoglobin in blood.

View Figure

Figure 2.6 Carbon dioxide dissociation curves for arterial blood (O2 Sat = 98%) and venous blood (O 2 Sat = 70%). The two points indicate the CO2 content of arterial and venous blood. The brackets show the relative contributions of hemoglobin desaturation (Haldane effect) and metabolic CO2 production (PcO2 effect) to the increase in CO2 content that occurs from arterial to venous blood. (From Forster RE II, DuBois A, Briscoe WA, et al. The lung, 3rd ed. Chicago: Yearbook Medical Publishers, 1986:238.)


The Haldane Effect Hemoglobin has a greater buffer capacity when it is in the desaturated form, and blood that is fully desaturated can bind an additional 60 mL/L of carbon dioxide. The increase in CO2 content that results from oxyhemoglobin desaturation is known as the Haldane effect. The CO2 dissociation curves in Figure 2.6 show that the Haldane effect plays an important role in the uptake of CO2 by venous blood. The two points on the graph show that the CO2 content in venous blood is 40 mL/L higher than in arterial blood. The brackets indicate that about 60% of the increased CO2 content in venous blood is due to an increase in PCO 2, while 40% is due to oxyhemoglobin desaturation. Thus, the Haldane effect is responsible for almost half of the rise in CO 2 content in venous blood. This is another example of the important role played by hemoglobin in CO 2 transport.

CO2 Elimination (VcO2) The dissociation of CO2 that occurs during transport in venous blood is reversed when the blood reaches the lungs. The reconstituted CO 2 is then eliminated through the lungs. The elimination of CO 2 (VcO 2) can P.36 be described by using an equation that is similar in form to the VO2 equation (Equation 2.8).

Figure 2.7 A schematic representation of the factors that contribute to CO2 elimination through the lungs (VCO2). The VCO2 is expressed as gas flow (mL/min) and as acid excretion (mEq/min). Q = cardiac output; CaCO2 = arterial CO2 content; CvCO2 = venous CO 2 content. View Figure

CvcO2 and CacO2 represent the CO 2 content in venous and arterial blood, respectively (note that the arterial and venous components are reversed when compared with the VO 2 equation). The determination of VcO2 by using the variables in Equation 2.14 is shown in Figure 2.7. Unfortunately, there are no simple derivative equations for CO 2 content in blood, so the VcO2 is usually measured directly. As shown in Table 2.3, the normal VcO2 in adults is 160–220 mL/min, or 90–130 mL/ min/m 2 when adjusted for body size. The VcO2 is normally about 80% of the VO2, so the VcO2/VO2 ratio is normally 0.8. The VcO 2/VO2 ratio, which is called the respiratory quotient (RQ), is used to identify the predominant type of nutrient substrate (i.e., protein, fat, or carbo-hydrate) being metabolized. Chapter 45 contains more information on the RQ.

VCO2 as Acid Excretion Carbon dioxide is essentially an acid because of its tendency to dissociate and form carbonic acid. Thus when the CO2 content is expressed in ion equivalents (mEq/L), the VcO2 (mEq/min) can be used to describe the rate of volatile acid excretion through the lungs. This is shown in P.37 Figure 2.7. The normal rate of acid excretion via the lungs is 9 mEq/min, or 12,960 mEq in 24 hours. Since the kidneys excrete only 40 to 80 mEq of acid every 24 hours (20), the principal organ of acid excretion in the body is the lungs, not the kidneys.

References General Works Halliwell B, Gutteridge JMC. Free radicals in biology and medicine, 3rd ed. New York: Oxford University Press, 1999.

Edwards JD, Shoemaker WC, Vincent J-L, eds. Oxygen transport: principles and practice. London: WB Saunders, 1993. Zander R, Mertzlufft F, eds. The oxygen status of arterial blood. Basel: Karger, 1991.

Oxygen Content in Blood 1. Pauling L. General chemistry, 3rd ed. Mineola, NY: Dover Publications, 1988:215.

2. Little RA, Edwards JD. Applied physiology. In: Edwards JD, Shoemaker WC, Vincent JL, eds. Oxygen transport: principles and practice. London: WB Saunders, 1993:21–40.

3. Zander R. Calculation of oxygen concentration. In: Zander R, Mertzlufft F, eds. The oxygen status of arterial blood. Basel: Karger, 1991:203–209.

4. Christoforides C, Laasberg L, Hedley-Whyte J. Effect of temperature on solubility of O2 in plasma. J Appl Physiol 1969;26:56–60. Bibliographic Links

Oxygen Delivery and Oxygen Uptake 5. Hameed SM, Aird WC, Cohn SM. Oxygen delivery. Crit Care Med 2003; 31(suppl):S658–S667. Ovid Full TextBibliographic Links 6. Little RA, Edwards JD. Applied physiology. In: Edwards JD, Shoemaker WC, Vincent J-L, eds. Oxygen transport: principles and practice. London, WB Saunders, 1993:21–40.

7. Bartlett RH. Oxygen kinetics: integrating hemodynamic, respiratory, and metabolic physiology. In: Critical care physiology. Boston: Little, Brown, 1996:1–23.

8. Ledingham IM, Naguib M. Overview: evolution of the concept from Fick to the present day. In: Edwards JD, Shoemaker WC, Vincent J-L, eds. Oxygen transport: principles and practice. London: WB Saunders, 1993:3–20.

9. Nunn JF. Nonrespiratory functions of the lung. In: Nunn JF, ed. Applied respiratory physiology. London: Butterworths 1993:306–317.

10. Jolliet P, Thorens JB, Nicod L, et al. Relationship between pulmonary oxygen consumption, lung inflammation, and calculated venous admixture in patients with acute lung injury. Intens Care Med 1996;22:277–285. Bibliographic Links 11. Leach RM, Treacher DF. The relationship between oxygen delivery and consumption. Dis Mon 1994;30:301–368. Bibliographic Links P.38 12. Connett RJ, Honig CR, Gayeski TEJ, et al. Defining hypoxia: a systems view of VO2, glycolysis, energetics, and intracellular PO 2. J Appl Physiol 1990;68: 833–842. Bibliographic Links 13. Shoemaker WC. Oxygen transport and oxygen metabolism in shock and critical illness. Crit Care Clin 1996;12:939–969. Bibliographic Links 14. Ronco J, Fenwick J, Tweedale M, et al. Identification of the critical oxygen delivery for anaerobic metabolism in critically ill septic and nonseptic humans. JAMA 1993;270:1724–1730. Ovid Full TextBibliographic Links 15. Lebarsky DA, Smith LR, Sladen RN, et al. Defining the relationship of oxygen delivery and consumption: use of biological system models. J Surg Res 1995;58:503–508. Full TextBibliographic Links

Carbon Dioxide Transport 16. Lambertson CJ. Carbon dioxide and respiration in acid-base homeostasis. Anesthesiology 1960;21:642–651. Bibliographic Links 17. Henneberg S, Soderberg D, Groth T, et al. Carbon dioxide production during mechanical ventilation. Crit Care Med 1987;15:8–13. Bibliographic Links

18. Brahm J. The red cell anion-transport system: kinetics and physiologic implications. In: Gunn R, Parker C, eds. Cell physiology of blood. New York: Rockefeller Press, 1988:142–150.

19. Nunn RF. Nunn's applied respiratory physiology, 4th ed. Oxford: Butterworth-Heinemann, 1993:220.

20. Forster RE II, DuBois A, Briscoe WA, et al. The lung, 3rd ed. Chicago: Yearbook Medical Publishers, 1986:223–247.

21. Comroe JH Jr. Physiology of respiration, 2nd ed. Chicago: Year Book, 1974: 201–210.

Preventive Practices in the Critically Ill We… repeatedly enlarge our instrumentalities without improving our purpose. --Will Durant

Chapter 3 Infection Control in the ICU Laymen always associate bacteria, microbes, and germs with disease. --John Postgate, Microbes and Man Microbial organisms (microbes) make up about 90% of the living matter on this planet. They're all around us: in the air we breathe, the food we eat, and the water we drink. They're on our skin, under our fingernails, in our nose and mouth, and armies of them congregate in our intestinal tract. Are these organisms the nasty little “germs” that are eager to invade the human body to conquer and destroy, as they are so often portrayed, or are they peace-loving creatures that mean us no harm? More the latter, it seems. Most microbes have nothing to gain by invading the human body (I'll exclude viruses here), but they have much to lose because they can be killed by the inflammatory response. It seems then that survival would dictate that microorganisms avoid the interior of the human body, not invade it. For more than a century, medicine has viewed the microbial world as an enemy that should be destroyed, and the practices described in this chapter are an expression of that belief. These practices are collectively known as “infection control,” and they are designed to prevent the spread of microorganisms from one person to another, or from one site to another on the same person. Most of the information in this chapter is taken from clinical practice guidelines published by the Centers for Disease Control and Prevention (CDC) and other expert agencies, and these are listed in the bibliography at the end of the chapter (1, 2, 3, 4, 5, 6, 7). As you will see, some infection control practices are rational, and some are ritual, but all are an essential part of daily life in the ICU. P.42

Skin Hygiene The surface of the skin is home to several species of bacteria and fungi, some of them attached to the underlying squamous cells of the skin (resident flora), and some of them are unattached and easily removed (transient flora) (3,4,8). Because most microbes are aquatic in nature and thrive in a moist environment, the microflora on the skin tend to congregate in moist regions like the groin and axilla. Contact surfaces like the skin on the hands can also be densely populated with microorganisms, and this microflora is a principal concern in infection control because it can be transmitted to others. An example of the organisms that populate the hands of ICU personnel is shown in Table 3.1. The most frequent isolate is Staphylococcus epidermidis (a coagulase-negative staphylococcus), followed by gram-negative enteric organisms and Candida species

(3,4,8,9). Eradicating microbes on the hands of hospital personnel is one of the holy crusades of infection control.

Cleaning vs Decontamination Plain soaps are detergents that can disperse particulate and organic matter, but they lack antimicrobial activity. Cleaning the skin with plain soap and water will remove dirt, soil, and organic matter from the skin, but will not eradicate the microbes on the skin. Scrubbing the skin with soap and water can remove transient (unattached) organisms, but the attached (resident) microorganisms are left in place. The removal of microbes from the skin, known as decontamination, requires the application of agents that have antimicrobial activity. Antimicrobial agents that are used to decontaminate the skin are called antiseptics, while those used to decontaminate inanimate objects are called disinfectants. TABLE 3.1 Organisms Isolated from the Hands of ICU Personnel


% Total Cultures

Gram-positive cocci Staph. epidermidis


Staph. aureus (MSSA)


Gram-negative Bacilli


Acinetobacter spp. Klebsiella spp. Enterobacter spp. Pseudomonas spp. Serratia spp. Yeasts and fungi Candida spp.


Candida spp. MSSA, methicillin-sensitive Staph. aureus. From Larson EL, Rackoff WR, Weiman M, et al. Assessment of two hand-hygiene regimens for intensive care unit personnel. Crit Care Med, 2001;29:944.

TABLE 3.2 Commonly Used Antiseptic Agents

Antiseptic Agent Alcohols



Rapid onset of action

Little residual activity

Broad spectrum of activity

Aqueous solutions can cause skin dryness.


Broad spectrum of activity

Slow onset of action Prolonged contact can irritate the skin


Good residual activity

Relatively narrow spectrum of activity An ocular irritant

From References 3, 4, and 8.


Antiseptic Agents The popular antiseptic agents in the United States are the alcohols (ethanol, propanol, and isopropyl alcohol), iodophors (slow-release iodine preparations), and chlorhexidine. (Hexachlorophene, once the most popular antiseptic agent in the U.S., is no longer recommended because of its limited spectrum of activity.) The relative advantages and disadvantages of each antiseptic agent are summarized in Table 3.2.

Alcohols The alcohols have excellent germicidal activity against gram-positive and gram-negative bacteria (including multidrug-resistant bacteria), various fungi (including Candida spp.), and viruses such as human immunodeficiency virus (HIV), hepatitis B virus (HBV), and hepatitis C virus (HCV) (3,4,8). Alcohol solutions containing 60% to 95% alcohol are most effective. Alcohols have a rapid onset of action but little persistent (residual) activity. They are less effective in the presence of dirt and organic matter, and are not recommended for use when the skin is visibly dirty or soiled with body fluids (e.g., blood) (4). Repeated use of aqueous (water-based) alcohol solutions can lead to drying and irritation of the skin, but these adverse effects are virtually eliminated when a waterless alcohol gel is used (4,8,9). Alcohol-impregnated towelettes are available but have limited amounts of alcohol and are no more effective in removing skin microbes than plain soap and water (4).

Iodophors Iodine is germicidal and has a broad spectrum of activity similar to the alcohols, but it is irritating to the skin and soft tissues. Skin irritation is reduced when a carrier molecule is used to release iodine slowly. Preparations that contain iodine and a carrier molecule are called iodophors, and the most popular iodophor in the United States is povidone-iodine (Betadine). Since the active ingredient in iodophors (iodine) is released slowly, iodophors must be left in contact with the skin for a few minutes to achieve maximal efficacy. However, prolonged contact with P.44 iodine can be irritating, so iodophors should be wiped from the skin after drying (3). Persistent (residual) activity is inconsistent after iodophors are wiped from the skin. Iodophors are neutralized by organic matter (3,4,9), so skin that is soiled with blood and body fluids should be cleaned before applying an iodophor. Povidone-iodine is usually provided as an aqueous solution, but alcohol-based solutions of povidone-iodine are available and may be more effective (10).

Chlorhexidine Chlorhexidine gluconate is a germicidal agent that is equally effective against gram-positive bacteria as the alcohols and iodophors, but is less effective against gram-negative bacilli and fungi. Its onset of action is slower than the alcohols but faster than the iodophors. The major advantage of chlorhexidine over the other antiseptic agents is its prolonged activity, which can last for six hours or longer (4). This is demonstrated in Figure 3.1. The residual activity is reduced by soaps and hand creams (4). Chlorhexidine is available in aqueous solutions ranging in strength from 0.5% to 4.0%. The 4% solution is most effective, but repeated use can P.45 cause skin irritation and dermatitis (4). Chlorhexidine is also an ocular irritant (4), and care should be taken to avoid contact with the eyes.

Figure 3.1 Comparative effects of a 6-minute hand scrub with 0.75% povidone-iodine (Betadine) and 4% chlorhexidine gluconate (Hibiclens) on microbial growth on the hands. Bacterial counts are expressed as log base 10 values. (From Peterson AF, Rosenberg A, Alatary SD, et al. Comparative evaluation of surgical scrub preparations. Surg Gynecol Obstet 1978;146:63*.Bibliographic Links)

View Figure

Spore-Forming Organisms None of the antiseptic agents described here is an effective sporicidal agent that can prevent the spread of spore-forming bacteria like Clos-tridium difficile and Bacillus anthracis (4). Gloves are needed whenever contact with these organisms is possible.

Handwashing Handwashing (a nebulous term that can include cleaning, antisepsis, or both) has been described as “…the single most important measure to reduce the risks of transmitting organisms from one person to another or from one site to another on the same patient” (ref. 2, updated guidelines). The recommendations for handwashing issued by the Centers for Disease Control are shown in Table 3.3. Note that an antiseptic solution rather than plain soap and water is recommended for most instances of handwashing, and that a waterless alcohol gel is recommended if the hands are not visibly soiled (remember that alcohol is much less effective in the presence of organic matter). The preference for alcohol gel is based P.46 on evidence that alcohol-containing products are superior to povidone-iodine or chlorhexidine solutions for reducing bacterial counts on the hands (4) and evidence that alcohol gels cause less skin irritation than antimicrobial soaps or aqueous antiseptic solutions (4,9). TABLE 3.3 Recommendations for Hand Hygiene

I. Handwashing with soap (plain or antiseptic) and water is recommended: 1. When hands are visibly dirty or contaminated with proteinaceous material or are visibly soiled with blood or other body fluids 2. Before eating 3. After leaving a restroom II. Handwashing with an antiseptic preparation a is recommended: 1. Before direct contact with patients 2. After contact with a patient's skin (intact or nonintact) 3. After contact with body fluids, secretions, excretions, mucous membranes, wound dressings, and contaminated items 4. Before donning sterile gloves to insert central intravascular catheters 5. Before inserting urinary catheters, peripheral venous catheters, or other invasive devices that do not require a surgical procedure 6. After removing gloves 7. When moving from a contaminated body site to a clean body site during patient care 8. After contact with inanimate objects in the immediate vicinity of the patient From References 2, 4, and 8. a A waterless alcohol gel is recommended if the hands are not visibly soiled. Otherwise, an antiseptic soap-and-water wash is recommended.

Compliance Despite the accolades showered on the practice of handwashing, surveys of ICU personnel reveal a consistent pattern of poor compliance with published guidelines for handwashing. Compliance rates are well below 50% in most surveys, and physicians are consistently the worst offenders (3,4,8,9). There are several reasons for this observation, and one of them is evident in Table 3.3: i.e., there are simply too many indications for handwashing. Anyone who has taken care of patients in an ICU will realize that full compliance with the recommendations in Table 3.3, particularly the recommendation that handwashing be performed before and after every patient contact, is neither practical, affordable, nor achievable on a consistent basis.

Technique Handwashing can be performed with plain soap or a variety of antiseptic preparations (soaps, aqueous solutions, or waterless gels). In general, alcohol-based products are more effective in reducing bacterial counts on the hands than are antiseptic soaps containing povidone-iodine or chlorhexidine (4). Whenever a soap (plain or antiseptic) is used, the wash should begin by wetting the hands with tap water. The soap should be applied to the palms of the hands and then rubbed over the entire surface of the hands for at least 30 seconds (4,8). Special attention should be given to the subungual areas under the fingernails, where microbes are usually most concentrated ( 3,4). The soap is then removed by rinsing with water, and the hands dried with a disposable towel. Hot water is not recommended for handwashing (4) because it is not more effective in

removing organisms from the skin than warm or cold water (11) and can be irritating to the skin. Using a disposable towel to dry the hands is equivalent to forced air drying (12) but is favored because it is quicker and more convenient. When a waterless alcohol gel is used, the hands should be cleaned first if necessary (remember that alcohol does not work well in the presence of organic matter), and the gel should be rubbed into the hands until they are dry. Repeated application of gels can leave the hands with a greasy feeling, and a periodic soap and water wash is sometimes preferred to remove any residual gel from the hands.

Protective Barriers Protective barriers like gloves, gowns, masks, and eye shields provide a physical impediment to the transmission of infectious agents. The principal role of these barriers is to protect hospital staff from infectious agents P.47 that can be transmitted by blood and body fluids, such as the human immunodeficiency virus (HIV) and hepatitis B and C viruses. TABLE 3.4 Recommendations for Glove Use in the ICU

I. Sterile gloves 1. Recommended for the following procedures A. Central venous catheterization B. Peripherally inserted central catheters (PICC) C. Artherial catheterization D. Placement of drainage catheters in a closed space (pleural, pericardial, or peritoneal cavities) E. Insertion of epidural catheters (for analgesia) or intraventricular catheters (for intracranial pressure monitoring II. Nonsterile gloves 1. Should be used for contact with any moist body substance—blood, body fluids, secretions, excretions, nonintact skin, and mucous membranes. Clean (unsoiled) gloves should be used for contact with nonintact skin and mucous membranes 2. Can be used for insertion of peripheral venous catheters as long as the gloved hands do not touch the catheter III. General recommendations 1. Gloves should be changed between tasks and procedures on the same patient if there has been contact with material that may be infectious 2. Gloves should be removed immediately after use, before contact with noncontaminated objects in the environment, and before going to another patient From References 2, 6, and 13.

Gloves Rubber gloves were popularized in this country in the late nineteenth century by William Halstead, the first (and enigmatic) Chief of the Surgery at Johns Hopkins Hospital, who covered only his palms and three fingers with the gloves because they were heavy and impaired the sense of touch. Today, sterile rubber gloves are the second skin of the operating surgeon. In the ICU, sterile gloves are used primarily for placing catheters in the bloodstream (see Table 3.4). In the 1980s (a century after the introduction of surgical gloves), the use of nonsterile gloves was popularized by the discovery that HIV is transmitted in blood and body fluids. This discovery prompted a policy known as Universal Precautions ( 1), which considered all patients as possible sources of HIV. An updated policy known as Standard Precautions (2,13) contains the current recommendations for nonsterile gloves, and these are shown in Table 3.4. Nonsterile gloves should be used for any contact with a moist body substance, which includes P.48 blood, body fluids, secretions, excretions, nonintact skin, and mucous membranes. Note also in Table 3.4 that nonsterile gloves are considered safe for insertion of peripheral venous catheters as long as a “no touch” technique is used (i.e., as long as the gloved hands are not permitted to touch the catheter) (6).

Handwashing and Gloves As indicated in Table 3.3, handwashing is recommended before donning gloves and again after they are removed. This recommendation is based on two concerns. The first is the fear that gloves can leak or tear and thereby allow microbial transmission between the hands of the healthcare worker and the patient. The second concern is the potential for moisture buildup on the hands during prolonged glove use, which would favor microbial growth on the hands while the gloves are on. Both of these are valid concerns for invasive surgical procedures, where glove use is prolonged and soiling of gloves is prominent. However, the significance of these concerns in a nonsurgical setting like the ICU (where glove use is not prolonged and soiling of gloves is usually not prominent) is less certain. The graph in Figure 3.2 provides some interesting observations about the need for antiseptic handwashing when gloves are used. The data in this graph is from a study involving two groups of ICU nurses: one group performed an antiseptic hand wash with 4% chlorhexidine before donning sterile gloves, while the other group did not wash their hands before donning gloves (14). Hand cultures were then obtained before, during, and after short-term glove use. The two graphs in Figure 3.2 show that microbial growth on the gloved hands was minimal in both groups, indicating that the pre-glove antiseptic handwash did not influence the infectious risk to patients from the gloved hands. The graphs also show that microbial activity on the hands was reduced in both groups after the gloves were removed. Thus, microbial proliferation on the hands is not a concern during short-term glove use. These results suggest that handwashing before and after short-term glove use in a nonsurgical setting like the ICU may be unnecessary.

Latex Allergy The dramatic increase in the use of rubber gloves over the last two decades has created a problem with latex hypersensitivity in hospital workers. Latex is a natural rubber product that is used in over 40,000 household and medical products, including gloves, face masks, blood pressure cuffs, and catheters (15). Repeated exposure to latex can

promote hypersensitivity reactions that can be evident clinically as either contact dermatitis (urticaria or eczema), anaphylaxis, rhinoconjunctivitis, or asthma ( 16,17). Latex hypersensitivity is reported in 10% to 20% of hospital workers, compared to 1% of the general population (16). For unclear reasons, patients with spina bifida have the highest risk of latex allergy, with as many as 40% of the population having this condition (18).

View Figure

Figure 3.2 Influence of pre-glove handwashing with an antiseptic agent (4% chlorhexidine) on hand cultures obtained during and after the use of sterile gloves. CFU = Colony-forming units 48 hours after the fingers of both hands were pressed directly on culture plates. The numbers in parentheses correspond to the values on the vertical axis of the graph. Note the break in the vertical axis and the different scales above and below the break. (From Rossoff LJ, Borenstein M, Isenberg HD. Is hand washing really needed in an intensive care unit? Crit Care Med 1995;23:1211, with permission.)


Diagnosis The diagnosis of latex allergy can be elusive. One problem is the nonspecific manifestations of disease. Another problem is the fact that symptoms of latex allergy can appear without direct physical contact with latex. This is often the case with the rhinoconjunctivitis and asthma, which are triggered by airborne latex particles. A history of symptoms confined to the workplace should create suspicion for latex allergy. The clinical manifestations of latex allergy often coincide with exposure to latex, so hospital workers with symptomatic latex allergy often display these symptoms while in the hospital and are symptom-free at home. There are two tests for latex hypersensitivity (19). One is a skin test, and the other is an assay for latex-specific immunoglobulin E levels in the bloodstream. Both have shortcomings. There is no standardized extract for the skin test (allergists have to make their own extract by pulverizing latex gloves!), so results are operator-dependent. The assay for latex-specific IgE in blood is currently the favored test, but the sensitivity can be low (19). If confronted with a case of possible latex allergy, you should contact the clinical laboratory in your hospital and ask about the availability and reliability of these tests in your region. P.50

Treatment The treatment of latex allergy is symptom-driven and nonspecific. Removing latex from the patient's immediate environment is the best strategy, but this may not be possible because latex is ubiquitous in the hospital environment (it is even found on tongue depressors!). The hospital should provide substitutes for latex products (e.g., vinyl gloves) when necessary.

Masks and Other Barriers As was the case with nonsterile gloves, the use of other barriers like masks, eye shields, face shields, and gowns increased markedly after the discovery that HIV is transmitted in blood and body fluids. These barriers are currently recommended for all procedures or patient care activities that are likely to generate splashes of blood, body fluids, secretions, or excretions (2,14). Nonsterile gowns are adequate, and gowns coated with a plastic covering are the least impervious to blood and body fluids (20). Soiled gowns and other barriers should be removed and discarded as soon as possible, and before going to another patient (14).

Types of Masks There are two types of face masks: surgical masks and respirators. Surgical masks were introduced to prevent contamination of the operative field during surgical procedures. In the past 2 decades, they have been adopted as a means of protecting healthcare workers from inhalation of airborne infectious agents. There is no evidence that surgical masks are effective in preventing infection (23), yet they continue to be used without question. Respirators are devices that protect the wearer from inhaling a dangerous substance (23). The different types of respirators include particulate respirators (block particulate matter), gas mask respirators (filter or clean chemical gases in the air), and the Self-Contained Breathing Apparatus (equipped with its own air tank), which is used by firefighters. Particulate respirators are used to block inhalation of airborne pathogens, particularly the tubercle bacillus that causes pulmonary tuberculosis. The respirator currently recommended for this purpose is called an N95 respirator ( 22,23). The “N” indicates that the mask will block non–oil-based or aqueous aerosols (the type that transmits the tubercle bacillus), and the “95” indicates the mask will block 95% of the intended particles (a requirement for a respirator mask to be judged effective) ( 23).

Types of Airborne Illness Infectious organisms that are capable of airborne transmission are divided into two categories: those greater than 5 microns (>;5m) in diameter, and those that are 5 microns or less (=5µ) in diameter. The organisms and airborne illnesses in each category are shown in Figure 3.3 (2). In each of these illnesses, airborne infectious particles are produced by coughing or sneezing (one cough or sneeze can produce 3,000 airborne P.51 particles) or procedures such as airways suctioning and bronchoscopy. The airborne particles can be inhaled or can impact on nonintact skin, or on the mucosa in the nose or mouth.

Figure 3.3 Infection control precautions for diseases that can spread via the airborne route. (From Reference 2.)

View Figure

Infectious particles >5µ in diameter usually travel no farther than 3 feet through the air, and to block transmission of these particles, a surgical mask is recommended (despite lack of proven efficacy!) when hospital staff or visitors are within 3 feet of the patient (2,21). The smaller (=5m in diameter) infectious particles can travel long distances in the air, and to prevent transmission of these particles, patients should be isolated in private rooms that are maintained at a negative pressure relative to the surrounding areas. For patients with infectious tuberculosis (pulmonary or laryngeal), all hospital staff and visitors should wear an N95 respirator P.52 mask while in the room (2,22). For patients in the infectious stages of rubeola (measles) and varicella (chickenpox or herpes zoster), individuals with no prior history of these infections who are also pregnant, immunocompromised, or otherwise debilitated by disease should not be allowed in the patient's room. For other susceptible individuals who must enter the room (i.e., hospital workers), an N95 respirator mask should be worn at all times while in the room.

Atypical Pulmonary TB It is important to distinguish infections caused by Mycobacterium tuberculosis from those caused by atypical mycobacteria (e.g., Mycobacterium avium complex) when determining the need for respiratory protection. Unlike the behavior of M. z, there is no evidence for person-to-person transmission of atypical mycobacteria (22), so special respiratory precautions (isolation and masks) are not required for patients with atypical pulmonary tuberculosis (2).

Blood-Borne Pathogens The greatest infectious risk you face in the ICU is exposure to blood-borne pathogens like HIV, hepatitis B virus (HBV), and hepatitis C virus (HCV). This section will describe the occupational risks associated with each of these pathogens and the preventive measures

used to minimize these risks.

Needlestick Injuries The transmission of blood-borne infections to hospital workers occurs primarily via needlestick injuries (i.e., accidental puncture wounds of the skin caused by hollow needles and suture needles). Each year, an estimated 10% of hospital workers sustain a needlestick injury (24). Most of these injuries occur in nurses, but the risk is also high in medical students, postgraduate trainees, and staff surgeons. As many as 70% of residents and medical students report a needlestick injury during their training (the incidence is highest in surgical residents) (25), and a survey in one hospital revealed that 60% of the staff surgeons experienced a needlestick injury at some time in their careers (26). The activities most often associated with needlestick injuries outside the operating room involve recapping and disposal of used needles (24).

Safety Devices The problem of needlestick injuries came to the attention of the United States Congress in the year 2000, and as a result, Congress passed the Needlestick Safety and Prevention Act that mandates the use of “safety-engineered” needles in all American health care facilities. The illustration in Figure 3.4 shows a simple safety device designed to eliminate the risk of needlestick injuries. The needle is equipped with a rigid, plastic housing that is attached by a hinge joint to the hub of the needle. The protective housing is normally positioned away from the needle so it P.53 does not interfere with needle use. When the needle is no longer needed, it is locked into the protective housing by holding the housing against a rigid structure and moving the needle about the hinge joint (like closing a door) until it snaps in place in the housing. The needle stays attached to the syringe during this procedure, and the hands never touch the needle. The protected needle and attached syringe are then placed in a puncture-proof “sharps container” for eventual disposal.

Figure 3.4 Safety device to prevent needlestick injuries from recapping and disposal of used needles.

View Figure

One-Handed Recapping Technique Once the needle is locked in its protective housing, it is not possible to remove it for further use. In situations where a needle has multiple uses (e.g., filling a syringe with a drug preparation and later injecting the drug P.54 in several increments), the needle can be rendered harmless between uses by recapping it with the one-handed “scoop technique” shown in Figure 3.5. With the needle cap resting on a horizontal surface, the needle is advanced into the needle cap. Using the tip of the needle cap as a fulcrum, the needle and cap are then lifted vertically until they are perpendicular to the horizontal surface. The needle is then pushed into the cap until it locks in place. The hands are never in a position to permit an accidental needle puncture.

Figure 3.5 The one-handed scoop technique for recapping a needle without the risk of a needlestick injury.

View Figure

Human Immunodeficiency Virus (HIV) The spread of HIV to hospital staff, although universally feared, is not a common event. As of June, 2000, there have been a total of 56 cases of HIV seroconversion in healthcare workers that can be definitely linked to HIV exposure in the workplace. Some of these cases involve laboratory P.55 workers, and only 44 cases involve percutaneous injury from hollow needles (the mode of transmission expected in ICUs) (24). Since HIV statistics were monitored for 15 years up to the year 2000, the 44 pertinent cases represent an average of 3 cases per year of HIV transmission in a nonoperative hospital setting. If all these cases occurred in the 6,000 ICUs in this country, the average yearly occurrence of HIV transmission in the ICU is one case for every 2,000 ICUs. Not much of a risk.

Percutaneous Exposures A needlestick puncture with a hollow needle will transfer an average of one microliter (10 -6 L) of blood (27). During the viremic stages of HIV infection, there are as many as 5 infectious particles per microliter of blood (28). Therefore puncture of the skin with a hollow needle that contains blood from a patient with active HIV infection is expected to transfer at least a few infectious particles. Fortunately, this is not enough to establish HIV infection in the recipient in most cases. A single needlestick injury with blood from an HIV-infected patient carries an average 0.3% risk of HIV seroconversion (5,24). This means that for every 1,000 needlestick injuries with HIV-infected blood, there will be an average of 3 cases of effective HIV transmission. The likelihood of HIV transmission is greater than 0.3% in the following circumstances: a deep skin puncture, visible blood on the needle, and injury from a needle that was placed in an artery or vein of the source patient (29).

Mucous Membrane Exposures Exposure of mucous membranes and nonintact skin to infectious body fluids carries much less risk of HIV transmission than needlestick injuries. A single exposure of broken skin or mucous membranes to blood from an HIV-infected patient carries an average 0.09% risk of HIV seroconversion (5,24). This means that for every 1,000 mucous membrane exposures to contaminated blood, there will be one (0.9) case of HIV transmission. (A one-in-a-thousand risk!)

Postexposure Management When a member of the ICU staff experiences a possible exposure to HIV from a needlestick injury or blood splash to the face, the appropriate steps to take are determined by the presence or absence of HIV anti-bodies in the blood of the source patient. If the HIV antibody status of the source patient is unknown, you should (with permission) perform a rapid HIV-antibody test on a blood sample from the source patient. This is done at the bedside (by an appropriately trained member of the staff), and the results are available in 10 to 15 minutes. The results of this test can be used to guide initial management decisions, but a positive result must be confirmed by another test such as a Western blot or immunofluorescent antibody assay. The recommendations for possible HIV exposure based on the HIV status of the source patient are outlined in Table 3.5 (5). The major decision following possible HIV transmission is whether or not to begin prophylactic therapy with antiretroviral agents in the exposed individual. If HIV infection is proven or suspected in the P.56 source patient, prophylactic therapy with at least 2 antiretroviral agents is started and continued for 4 weeks (or until there is convincing evidence for the absence of HIV infection in the source patient). A popular two-drug regimen is the combination of zidovudine (200 mg TID) and lamovudine (150 mg daily). These two drugs are available in a combination tablet (COMBIVIR, GlaxoSmithKline), each containing 150 mg lamovudine and 300 mg zidovudine. A third antiretroviral agent is added if there is evidence for symptomatic or advanced HIV infection in the source patient, or if the HIV exposure is severe (i.e., deep needlestick injury, injury from a needle soiled with infectious blood, or injury from a needle that was placed in an artery or vein of an HIV-infected patient) (29). The agents that can be added to the two-drug regimen are included at the bottom of Table 3.5. TABLE 3.5 Postexposure Prophylaxis for HIV Infection

Indications for Each Type of Postexposure Drug Regimen No Drugs 1. When source is HIV-negative 2. When HIV status of source is not known but HIV is unlikelyc 3. When source is not known but HIV is unlikelyd

Two Drugs a 1. When source is HIV-positive but asymptomatic 2. When HIV status of source is not known but HIV is likely.c 3. When source is not known but HIV is likelyd

Three Drugsb 1. When source is HIV-positive and symptomatic 2. When source is HIV-positive and asymptomatic but exposure is severee

From Reference 5. a The recommended two-drug regimen is zidovudine (200 mg TID) plus lamivudine (150 mg BID) for 4 weeks. The two agents are available together as COMBIVIR. b Add one of the following drugs to the two-drug regimen: efavirenz (600 mg at bedtime), indinavir (800 mg every 8 hr, between meals), or nelfinavir (2.5 g daily in 2 or 3 divided doses, with meals). c When the HIV status of the source is unknown, the likelihood of HIV is based on the presence or absence of risk factors. d When the source is unknown, the likelihood of HIV is based on the prevalence of HIV in the population served. e Severe exposure is defined as deep injury, needle soiled with blood from source patient, and exposure from needle inserted into artery or vein of source patient.

It is important to emphasize that the current recommendations for prophylaxis with antiretroviral agents are empiric, and not based on proven efficacy. Even if antiretroviral therapy is completely effective in preventing HIV transmission, an average of 330 patients who have been exposed to HIV-infected blood would have to be treated to prevent one case of HIV transmission. Considering that the prophylactic regimens of antiretroviral drug therapy are poorly tolerated (one of every three subjects given antiretroviral drugs for postexposure prophylaxis will stop taking the drugs because of troublesome side effects) (5), the risks of antiretroviral P.57 drug prophylaxis may outweigh the overall benefit in many subjects, particularly when HIV infection in the source patient is not proven.

Postexposure Surveillance

Antibody responses to acute or primary HIV infection can take 4 to 6 weeks to become evident. Therefore anyone with documented exposure to HIV infection should have serial tests for HIV antibodies at 6 weeks, 3 months, and 6 months after the exposure (5). More prolonged testing is not warranted unless the exposed person develops symptoms compatible with HIV infection.

Postexposure Hotline The National Clinicians' Postexposure Prophy-laxis Hotline (PEP line) is a resource for anyone with questions about postexposure prophylaxis for HIV infection. The toll-free number is 888-448-4911.

Hepatitis B Virus The blood-borne hepatitis B virus (HBV) is much more transmission-prone than HIV. One microliter (10-6 L) of blood from a patient with HBV-induced acute hepatitis can have as many as one million infectious particles, whereas, as just mentioned, a similar volume of blood from a patient with active HIV infection will have only 5 or fewer infectious particles (28). Fortunately, there is a vaccine that can provide immunity against HBV infection.

Hepatitis B Vaccination Vaccination against hepatitis B is recommended for anyone who has contact with blood, body fluids, and sharp instruments (5), which is virtually everyone who works in an ICU. The only contraindication to the vaccine is a prior history of anaphylaxis from baker's yeast (5). The vaccination involves 3 doses and should proceed as follows (5): 1. The first 2 doses (given by deep IM injection) are given 4 weeks apart, and the third dose is administered 5 months after the second dose. 2. If the vaccination series is interrupted after the first dose, the whole sequence is not repeated. If the second dose was missed, it is given as soon as possible, and the third dose is administered at least 2 months later. If the third dose was missed, it is administered as soon as possible, and the vaccination series is considered completed. The hepatitis B vaccine produces immunity by stimulating production of an antibody to the hepatitis B surface antigen (anti-HBs). The primary vaccination series is not always successful in providing immunity, so the following evaluation is recommended (5). 3. One to two months after the vaccination is completed, the serum anti-HBs level should be measured. Immunity is indicated by an anti-HBs level that is =10 mIU/mL. If the anti-HBs is 032% of infusions (31).

The Colloid–Crystalloid Wars There is a long standing (and possibly eternal) debate concerning the type of fluid (crystalloid or colloid) that is most appropriate for volume resuscitation. Each fluid has its army of loyalists who passionately defend the merits of their fluid. The following are the issues involved in this debate.

Early Focus on Crystalloids Early studies of acute blood loss in the 1960s produced two observations that led to the popularity of crystalloid fluids for volume resuscitation. The first observation was a human study showing that acute blood loss is accompanied by a shift of interstitial fluid into the bloodstream (transcapillary refill), leaving an interstitial fluid deficit (34). The second

observation was an animal study showing that survival in hemorrhagic shock is improved if crystalloid fluid is added to reinfusion of the shed blood (35). The interpretation of these two observations, at that time, was that a major consequence of acute blood loss was an interstitial fluid deficit P.247 and that replenishing this deficit with a crystalloid fluid will reduce mortality. Thus crystalloid fluids were popularized for volume resuscitation because of their ability to add volume to the interstitial fluids. Later studies using more sensitive measures of interstitial fluid revealed that the interstitial fluid deficit in acute blood loss is small and is unlikely to play a major role in determining the outcome from acute hemorrhage. This refuted the importance of filling the interstitial fluid compartment with crystalloids, yet the popularity of crystalloid fluids for volume resuscitation did not wane.

The Goal of Volume Resuscitation The most convincing argument in favor of colloids for volume resuscitation is their superiority over crystalloid fluids for expanding the plasma volume. Colloid fluids will achieve a given increment in plasma volume with only one-quarter to one-third the volume required of crystalloid fluids. This is an important consideration in patients with brisk bleeding or severe hypovolemia, where rapid volume resuscitation is desirable. The proponents of crystalloid resuscitation claim that crystalloids can achieve the same increment in plasma volume as colloids. This is certainly the case, but three to four times more volume is required with crystalloids than colloids to achieve this goal. This adds fluid to the interstitial space and can produce unwanted edema. In fact, as mentioned earlier (and demonstrated in Figure 13.2), the principal effect of crystalloid infusions is to expand the interstitial fluid volume, not the plasma volume. Since the goal of volume resuscitation is to support the intravascular volume, colloid fluids are the logical choice over crystalloid fluids.

Filling a Bucket The following example illustrates the problem with using crystalloids to expand the plasma volume. Assume that you have two buckets, each representing the intravascular compartment, and each bucket is connected by a clamped hose to an overhanging reservoir that contains fluid. One reservoir contains a colloid fluid in the same volume as the bucket, and the other reservoir contains a crystalloid fluid in a volume that is three to four times greater than the colloid volume. Now release the clamp on each hose and empty the reservoirs: both buckets will fill with fluid, but most of the crystalloid fluid will spill over onto the floor. Now ask yourself which method is better suited for filling buckets: the colloid method, with the right amount of fluid and no spillage, or the crystalloid method, with too much fluid, most of which spills onto the floor.

Clinical Outcome As mentioned earlier (see section on Safety of Albumin Solutions), the bulk of available evidence indicates that neither type of resuscitation P.248 fluid provides a survival benefit (24,25,26), while colloid (albumin-containing) fluids may cause fewer adverse events (27).

The Problem with Mortality Studies There are two problems with the studies comparing mortality rates associated with colloid and crystalloid fluids. The first problem is that most studies included a diverse group of patients who could have died from a variety of illnesses, and there is no way of determining if an intravenous fluid was directly related to the cause of death. For example, a resuscitation fluid could restore a normal plasma volume, but the patient dies of pneumonia: in this case, the fluid should not be blamed for the death. The second problem is the assumption that an intervention must save lives to be considered beneficial. It seems that an intervention should be judged by whether it achieves its intended goal (e.g., a resuscitation fluid should be judged by how well it restores plasma volume); determining if that goal influences mortality is a separate question.

Expense The biggest disadvantage of colloid resuscitation is the higher cost of colloid fluids. Table 13.4 shows a cost comparison for colloid and crystalloid fluids. Using equivalent volumes of 250 mL for colloid fluids and 1,000 mL for crystalloid fluids, the cost of colloid resuscitation is nine times higher (if hetastarch is used) to twenty-one times higher (if albumin is used) than volume resuscitation with crystalloid fluids.

A Suggestion Most studies comparing colloid and crystalloid fluids have attempted to determine if one type of resuscitation fluid is better than the other for all critically ill patients. This seems unreasonable, considering the multitude P.249 of clinical problems encountered in ICU patients. A more reasonable approach would be to determine if one type of fluid is more appropriate than the other for a given clinical condition (36). For example, patients with hypovolemia secondary to dehydration (where there is a uniform loss of extracellular fluid) might benefit more from a crystalloid fluid (which is expected to fill the extracellular space uniformly) than a colloid fluid, and patients with hypovolemia secondary to hypoalbuminemia (where there are fluid shifts from the intravascular to extravascular space) might benefit more from a colloid fluid (particularly 25% albumin) than a crystalloid fluid. Tailoring the type of resuscitation fluid to the specific clinical condition seems a more logical approach than using the same type of fluid without exception for all ICU patients. TABLE 13.4 Relative Cost of Intravenous Fluids



Unit size


Crystalloid fluids Isotonic saline


1,000 ml


Lactated Ringer's


1,000 ml


5% Albumin


250 ml


25% Albumin


50 ml


6% Hetastarch


500 ml


6% Dextran-70


500 ml


Colloidal fluids


Average wholesale price listed in 2005 Redbook. Montvale, NJ: Thomson PDR, 2005.

Hypertonic Resuscitation Volume resuscitation with hypertonic saline (7.5% NaCl) has received much attention as a method of small-volume resuscitation. A 7.5% P.250 sodium chloride solution has an osmolality that is about 8.5 times greater than plasma (see Table 13.1). Figure 13.2 demonstrates that infusion of 250 mL of 7.5% NaCl will increase plasma volume by about twice the infused volume, indicating that hypertonic saline allows for volume resuscitation with relatively small volumes. Also note in Figure 13.2 that the total increase in extracellular fluid volume (1,235 mL) produced by 7.5% NaCl is about 5 times greater than the infused volume (250 mL). The additional volume comes from intracellular fluid that moves out of cells and into the extracellular space. This demonstrates one of the feared complications of hypertonic saline resuscitation: cell dehydration.

Figure 13.5 A comparison of the volume of three intravenous fluids needed to maintain a normal rate of aortic blood flow in an animal model of hemorrhagic shock. (From Chiara O, Pelosi P, Brazzi L, et al. Resuscitation from hemorrhagic shock: experimental model comparing normal saline, dextran, and hypertonic saline solutions. Crit Care Med 2003;31:1915.Ovid Full TextBibliographic Links) View Figure

What Role? The small volumes required with hypertonic saline resuscitation have been proposed as a possible benefit in the resuscitation of trauma victims with head injuries (to limit the severity of cerebral edema). However, the effective resuscitation volumes with hypertonic saline are similar to colloid resuscitation, as shown in Figure 13.5 (37), and a recent clinical study documented no advantage with hypertonic saline in the prehospital resuscitation of patients with traumatic head injury (38). At the present time, hypertonic saline is a resuscitation fluid without a clear indication.

A Final Word There is too much chatter about which type of resuscitation fluid (colloid or crystalloid) is most appropriate in critically ill patients because it is unlikely that one type of fluid is best for all patients. A more logical approach is to select the type of fluid that is best designed to correct a specific problem with fluid balance. For example, crystalloid fluids are designed to fill the extracellular space (interstitial space plus intravascular space) and would be appropriate for use in patients with dehydration (where there is a loss of interstitial fluid and intravascular fluid). Colloid fluids are designed to expand the plasma volume and are appropriate for patients with hypovolemia due to blood loss, while albumin-containing colloid fluids are appropriate for patients with hypovolemia associated with hypoalbuminemia. Tailoring fluid therapy to specific problems of fluid imbalance is the best approach to volume resuscitation in the ICU.

References Reviews 01. Alderson P, Schierhout G, Roberts I, et al. Colloid versus crystalloids for fluid

resuscitation in critically ill patients: a systematic review of randomized trials. Cochrane Database Syst Rev 2000; Issue 3, Art. No. CD000567. 02. Choi PT, Yip G, Quinonez LG, et al. Crystalloids vs. colloids in fluid resuscitation: a systematic review. Crit Care Med 1999;27:200–210.Ovid Full TextBibliographic Links P.251 03. Whinney RR, Cohn SM, Zacur SJ. Fluid resuscitation for trauma patients in the 21st century. Curr Opin Crit Care 2000;6:396–400. 04. Bunn F, Alderson P, Hawkins V. Colloid solutions for fluid resuscitation. Cochrane Database Syst Rev 2001;:CD001319.Bibliographic Links

Crystalloid Fluids 1. Imm A, Carlson RW. Fluid resuscitation in circulatory shock. Crit Care Clin 1993;9:313–333. Bibliographic Links 2. Scheingraber S, Rehm M, Schmisch C, et al. Rapid saline infusion produces hyperchloremic acidosis in patients undergoing gynecologic surgery. Anesthesiology 1999;90:1265–1270. Ovid Full TextBibliographic Links 3. Prough DS, Bidani A. Hyperchloremic metabolic acidosis is a predictable consequence of intraoperative infusion of 0.9% saline. Anesthesiology 1999;90:1247–1249. Ovid Full TextBibliographic Links 4. Griffith CA. The family of Ringer's solutions. J Natl Intravenous Ther Assoc 1986;9:480–483.

5. American Association of Blood Banks. Technical manual, 10th ed. Arlington, VA: American Association of Blood Banks, 1990:368.

6. King WH, Patten ED, Bee DE. An in vitro evaluation of ionized calcium levels and clotting in red blood cells diluted with lactated Ringer's solution. Anesthesiology 1988;68:115–121. Bibliographic Links 7. Didwania A, Miller J, Kassel; D, et al. Effect of intravenous lactated Ringer's

solution infusion on the circulating lactate concentration, part 3: result of a prospective, randomized, double-blind, placebo-controlled trial. Crit Care Med 1997;25:1851–1854. Ovid Full TextBibliographic Links 8. Jackson EV Jr, Wiese J, Sigal B, et al. Effects of crystalloid solutions on circulating lactate concentrations, part 1: implications for the proper handling of blood specimens obtained from critically ill patients. Crit Care Med 1997;25: 1840–1846. Ovid Full TextBibliographic Links 9. Halpern NA, Alicea M, Seabrook B, et al. [Q]olyte S, a physiologic multielectrolyte solution, is preferable to normal saline to wash cell saver salvaged blood: conclusions from a prospective, randomized study in a canine model. Crit Care Med 1997;12:2031–2038. Ovid Full TextBibliographic Links

Dextrose Solutions 10. Gunther B, Jauch W, Hartl W, et al. Low-dose glucose infusion in patients who have undergone surgery. Arch Surg 1987;122:765–771. Bibliographic Links 11. DeGoute CS, Ray MJ, Manchon M, et al. Intraoperative glucose infusion and blood lactate: endocrine and metabolic relationships during abdominal aortic surgery. Anesthesiology 1989;71:355–361. Bibliographic Links 12. Turina M, Fry D, Polk HC Jr. Acute hyperglycemia and the innate immune system: clinical, cellular, and molecular aspects. Crit Care Med 2005;33:1624–1633. Ovid Full TextBibliographic Links 13. Van Den Berghe G, Wouters P, Weekers F, et al. Intensive insulin therapy in critically ill patients. N Engl J Med 2001;345:1359–1367. Ovid Full TextBibliographic Links 14. Sieber FE, Traystman RJ. Special issues: glucose and the brain. Crit Care Med 1992;20:104–114. Bibliographic Links P.252 15. Finney SJ, Zekveld C, Elia A, et al. Glucose control and mortality in critically ill patients. JAMA 2003;290:2041–2047. Ovid Full TextBibliographic Links

Colloid Fluids 16. Jacob G, Raj S, Ketch T, et al. Postural pseudoanemia: posture-dependent changes in hematocrit. Mayo Clin Proc 2005;80:611–614. Full TextBibliographic Links 17. Griffel MI, Kaufman BS. Pharmacology of colloids and crystalloids. Crit Care Clin 1992;8:235–254. Bibliographic Links 18. Kaminski MV, Haase TJ. Albumin and colloid osmotic pressure: implications for fluid resuscitation. Crit Care Clin 1992;8:311–322. Bibliographic Links 19. Sutin KM, Ruskin KJ, Kaufman BS. Intravenous fluid therapy in neurologic injury. Crit Care Clin 1992;8:367–408. Bibliographic Links 20. Imm A, Carlson RW. Fluid resuscitation in circulatory shock. Crit Care Clin 1993;9:313–333. Bibliographic Links 21. Soni N, Margarson M. Albumin, where are we now? Curr Anesth Crit Care 2004;15:61–68.

22. Halliwell B. Albumin: an important extracellular antioxidant? Biochem Pharmacol 1988;37:569–571.

23. Cochrane Injuries Group Albumin Reviewers. Human albumin administration in critically ill patients: systematic review of randomized, controlled trials. Br Med J 1998;317:235–240. Bibliographic Links 24. Choi PT-L, Yip G, Quinonez LG, et al. Crystalloids vs. colloids in fluid resuscitation: a systematic review. Crit Care Med 1999;27:200–210. Ovid Full TextBibliographic Links 25. Wilkes MN, Navickis RJ. Patient survival after human albumin administration: a meta-analysis of randomized, controlled trials. Ann Intern Med 2001;135: 149–164. Ovid Full TextFull TextBibliographic Links 26. SAFE Study Investigators. A comparison of albumin and saline for fluid

resuscitation in the intensive care unit. N Engl J Med 2004;350:2247–2256. Ovid Full TextBibliographic Links 27. Vincent J-L, Navickis RJ, Wilkes MM. Morbidity in hospitalized patients receiving human serum albumin: a meta-analysis of randomized, controlled trials. Crit Care Med 2004;32:2029–2038. Ovid Full TextBibliographic Links 28. Treib J, Baron JF, Grauer MT, et al. An international view of hydroxyethyl starches, Intensive Care Med 1999;25:258–268.

29. de Jonge E, Levi M. Effects of different plasma substitutes on blood coagulation: a comparative review. Crit Care Med 2001;29:1261–1267. Ovid Full TextBibliographic Links 30. Haynes GR, Havidich JE, Payne KJ. Why the Food and Drug Administration changed the warning label for hetastarch. Anesthesiology 2004;101:560–561. Ovid Full TextBibliographic Links 31. Nearman HS, Herman ML. Toxic effects of colloids in the intensive care unit. Crit Care Clin 1991;7:713–723. Bibliographic Links 32. Gan TJ. Hextend, a physiologically balanced plasma expander for large volume use in major surgery: a randomized phase III clinical trial. Anesth Analg 1999;88:992–998. Ovid Full TextBibliographic Links 33. Drumi W, Polzleitner D, Laggner AN, et al. Dextran-40, acute renal failure, and elevated plasma oncotic pressure. N Engl J Med 1988;318:252–254. Bibliographic Links

Colloid–Crystalloid Wars 34. Moore FD. The effects of hemorrhage on body composition. N Engl J Med 1965;273:567–577. Bibliographic Links P.253 35. Shires T, Carrico J, Lightfoot S. Fluid therapy in hemorrhagic shock. Arch Surg 1964;88:688–693. Bibliographic Links

36. Weil MH, Tang W. Albumin versus crystalloid solutions for the critically ill and injured [Editorial]. Crit Care Med 2004;32:2154–2155. Ovid Full TextBibliographic Links

Hypertonic Resuscitation 37. Chiara O, Pelosi P, Brazzi L, et al. Resuscitation from hemorrhagic shock: experimental model comparing normal saline, dextran, and hypertonic saline solutions. Crit Care Med 2003;31:1915–1922. Ovid Full TextBibliographic Links 38. Cooper DJ, Myles PS, McDermott FT, et al. Prehospital hypertonic saline resuscitation of patients with hypotension and severe traumatic brain injury. JAMA 2004;291:1350–1357. Ovid Full TextBibliographic Links

Chapter 14 Acute Heart Failure Syndromes There's no doubt that the proper functioning of our pipes and pumps does have an immediate urgency well beyond that of almost any of our other bits and pieces. --Steven Vogel (Vital Circuits, 1992) Acute or decompensated heart failure is responsible for about 1 million hospital admissions each year in the United States (1), and it is the leading cause of hospital admissions for adults over the age of 65 (2). Heart failure is not a single entity but can be classified according to the side of the heart that is involved (right-sided vs. left-sided failure) or the portion of the cardiac cycle that is affected (diastolic vs. systolic failure). This chapter describes the diagnostic and therapeutic approach to each of these four heart failure syndromes using the principles of cardiac performance described in Chapter 1 (2, 3, 4, 5, 6). The approach to heart failure in this chapter is designed for the ICU: it is based on invasive hemodynamic measurements, rather than clinical symptoms and signs, and focuses on the mechanical problems of heart failure rather than the responsible diseases. The usual causes of heart failure are shown in Figure 14.1.

Hemodynamic Alterations The hemodynamic consequences of progressive left-sided heart failure are shown in Figure 14.2. (The measurements in this graph were obtained from a patient who had just undergone cardiopulmonary bypass surgery). The hemodynamic changes progress through three stages (the numbers below correspond to the circled numbers in Figure 14.2): 1. The earliest sign of ventricular dysfunction is an increase in cardiac filling pressures. The stroke volume is maintained, but at the expense of the elevated filling pressure.

Figure 14.1 Common causes of acute heart failure, listed according to the anatomic region involved. RV = right ventricle; LV = left ventricle.

View Figure

P.256 2. The next stage is marked by a decrease in stroke volume and an increase in heart rate. The tachycardia offsets the reduction in stroke volume, so the cardiac output remains unchanged. 3. The final stage is characterized by a decrease in cardiac output. The point at which the cardiac output begins to decline marks the transition from compensated to decompensated heart failure. The serial hemodynamic changes shown in Figure 14.2 demonstrate that cardiac output is impaired only in the more advanced stages of heart failure; therefore a normal cardiac output is not necessarily a normal cardiac pump. Cardiac pump function should be evaluated using the relationship between ventricular filling pressure and stroke volume. This relationship is the basis for ventricular function curves, which are described in Chapter 1 (see Figure 1.2).

Systolic Versus Diastolic Failure Heart failure is not synonymous with contractile failure because systolic function is normal in 40 to 50% of newly-diagnosed cases of heart P.257 failure (2). The problem in this condition is a combination of impaired ventricular relaxation and a decrease in passive ventricular distensibility, a disorder known as diastolic heart failure (6,7). In this type of heart failure, the decrease in cardiac output is due to inadequate ventricular filling, not impaired systolic contraction. Common causes of diastolic heart failure in ICU patients include ventricular hypertrophy, myocardial ischemia (stunned myocardium), and positive-pressure mechanical ventilation.

Figure 14.2 Hemodynamic effects of progressive left-sided heart failure in a postoperative patient.

View Figure

Diagnostic Difficulties The usual method of evaluating cardiac pump function (by the relationship between ventricular filling pressure and stroke volume) will not distinguish between diastolic and systolic heart failure (7,8). This is illustrated in Figure 14.3. The curves in this figure are similar to the pressure–volume curves shown in Figures 1.2 and 1.3. The upper curves in the figure are ventricular function curves relating ventricular end-diastolic pressure and stroke volume. These curves indicate that heart failure is associated with an increase in end-diastolic pressure and a decrease in stroke volume. It is not possible, however, to determine if P.258 the heart failure is systolic or diastolic based on these measurements. The lower set of curves shows the pressure–volume relationships during diastole in the two types of heart failure. The end-diastolic pressure is increased in both types of heart failure, but the end-diastolic volume changes in different directions: it is increased in systolic heart failure and decreased in diastolic heart failure. Thus the end-diastolic volume, not the end-diastolic pressure, is the hemodynamic measure that will distinguish diastolic from systolic heart failure. There is a specialized pulmonary artery catheter that measures the end-diastolic volume of the right ventricle (see later), but otherwise this measurement is not readily available.

Figure 14.3 Graphs showing diastolic pressure volume curves in systolic and diastolic heart failure (lower curves) and ventricular function curves in heart failure (upper curves). The ventricular function curves, which are used to evaluate cardiac function in the clinical setting, are unable to distinguish between diastolic and systolic failure.

View Figure

Ventricular Ejection Fraction The measurement that is most often used to distinguish between diastolic and systolic heart failure is the ventricular ejection fraction (EF), which is a measure of the strength of ventricular contraction. The EF expresses the stroke volume (SV) as a fraction of the end-diastolic volume (EDV) P.259 The normal EF of the right ventricle is 0.50 to 0.55, and the normal EF of the left ventricle is 0.40 to 0.50. The EF is normal in patients with diastolic heart failure and is reduced in patients with systolic heart failure. Cardiac ultrasound can be used to measure ventricular EF at the bedside. Transthoracic ultrasound can be used to measure the EF of the left ventricle (6,7), and transesophageal ultrasound can be used to measure the EF of the right ventricle (8). A specialized pulmonary artery catheter is also available for measuring the EF of the right ventricle, as described in the next section.

Right Versus Left Heart Failure Right heart failure (which is predominantly systolic heart failure) is more prevalent than suspected in ICU patients (9), and it may be particularly prominent in ventilator-dependent patients. The following measurements can prove useful in identifying right heart failure.

Cardiac Filling Pressures The relationship between the central venous pressure (CVP) and the pulmonary capillary wedge pressure (PCWP) can sometimes be useful for identifying right heart failure. The following criteria have been proposed for right heart failure ( 10): CVP > 15 mm Hg and CVP = PCWP or CVP > PCWP. Unfortunately, at least one-third of patients with acute

right heart failure do not satisfy these criteria (10). One problem is the insensitivity of the CVP; an increase in the CVP is seen only in the later stages of right heart failure. Contractile failure of the right ventricle results in an increase in end-diastolic volume, and only when the increase in volume of the right heart is impeded by the pericardium does the end-diastolic pressure (CVP) rise (9). Another problem with the CVP–PCWP relationship for identifying right heart failure is the interaction between the right and left sides of the heart. This is shown in Figure 14.4. Both ventricles share the same septum, so enlargement of the right ventricle pushes the septum to the left and compromises the left-ventricular chamber. This interaction between right and left ventricles is called interventricular interdependence, and it can confuse the interpretation of ventricular filling pressures. In fact, as indicated by the diastolic pressures in Figure 14.4, the hemodynamic changes in right heart failure can look much like the hemodynamic changes in pericardial tamponade (9).

Thermodilution Ejection Fraction A specialized pulmonary artery catheter is available that uses a fast-response thermistor to measure the ejection fraction (EF) of the right ventricle (11). Rapid-response thermistors can record the temperature changes associated with each cardiac cycle. This produces a thermodilution curve like the one shown in Figure 14.5. The change in temperature between each plateau on the curve is caused by dilution of the cold indicator fluid by venous blood that fills the ventricle during diastole. Because the volume that fills the ventricles during diastole is equivalent P.260 to the stroke volume, the temperature difference T1 - T2 is the thermal equivalent of the stroke volume (SV), and the temperature T 1 is thus a thermal marker of the end-diastolic volume (EDV). The ejection fraction is then equivalent to the ratio T 1 - T2/T1 (or SV/EDV). Once the EF is measured, the stroke volume can be measured in the usual fashion (as cardiac output divided by heart rate), and the EDV can be determined by rearranging Equation 14.1. Figure 14.4 Interventricular interdependence: the mechanism whereby right heart failure can reduce diastolic filling of the left ventricle and increase the left-ventricular end-diastolic (wedge) pressure. RV = right ventricle; LV = left ventricle. The numbers in each chamber represent the systolic pressure as the numerator and the end-diastolic pressure as the denominator.

View Figure

The normal right ventricular EF (RVEF) using thermodilution is 0.45 to 0.50, which is about 10% lower than the EF measured by radionuclide imaging (the gold standard). The normal right ventricular EDV (RVEDV) is 80 to 140 mL/m 2. Figure 14.5 The thermodilution method of measuring the ejection fraction (EF) of the right ventricle using thermal equivalents for end-diastolic volume (EDV), end-systolic volume (ESV), and stroke volume (SV). View Figure

P.261 Since most cases of right heart failure represent systolic failure, the RVEF is expected to be less than 0.45, and the RVEDV is expected to be above 140 mL/m2 in cases of right heart failure (12). The response of RVEDV to a fluid challenge may also be diagnostic: volume infusion is expected to increase the RVEDV in patients with right heart failure, while in other patients, the RVEDV is unchanged after a fluid challenge (13).

Echocardiography Cardiac ultrasound can be useful at the bedside for differentiating right from left heart failure. Three findings typical of right heart failure are (a) an increase in right-ventricular chamber size, (b) segmental wall motion abnormalities on the right, and (c) paradoxical motion of the interventricular septum (10).

B-Type Natriuretic Peptide Brain-type (B-type) natriuretic peptide (BNP) is a neurohormone that is released by the ventricular myocardium in response to ventricular volume and pressure overload. Plasma levels of BNP increase in direct relation to increases in ventricular end-diastolic volume and end-diastolic pressure (both right-sided and left-sided), and the rise in BNP produces both vasodilatation and an increase in renal sodium excretion (14).

Diagnostic Value The plasma BNP level has proven to be an important tool for the diagnosis of heart failure. In patients who present with dyspnea of unknown etiology, a plasma BNP > 100 picograms/milliliter (pg/mL) can be used P.262 as evidence of heart failure as a cause of the dyspnea (diagnostic accuracy = 84%)

(15). In fact, in patients who present to the emergency department with dyspnea of unknown etiology, the plasma BNP level (using a cutoff level of 100 pg/mL) is the single most accurate predictor of the presence or absence of heart failure (15). Rapid determination of plasma BNP levels is available at the bedside using a fluorescence immunoassay kit (Triage; Biosite Diagnostics, San Diego, CA) that allows for timely identification of acute heart failure in the emergency department (15). Plasma BNP levels also show a direct correlation with the severity of heart failure (14,16) [i.e., plasma levels are higher in patients with more advanced stages of heart failure (see Table 14.1)]. This correlation indicates that plasma BNP levels may be useful for monitoring the clinical course of heart failure.

Other Contributing Factors Plasma BNP levels are influenced by gender, age, and renal function. This is demonstrated in Table 14.1. Plasma BNP levels are about 50% higher in females than in males, and plasma levels increase with advancing age in both sexes ( 14). Renal insufficiency also increases plasma BNP levels (because BNP is cleared by the kidneys), but levels usually do not pass the 100 pg/mL threshold unless there is associated volume overload (see Table 14.1) (17). TABLE 14.1 Plasma BNP Levels in Selected Conditions


Mean Plasma BNP (pg/mL)

Females—no CHFa Age 55–64


Age 75+


Males—no CHFa Age 55–64


Age 75+


Renal insufficiency b No volume overload Volume overload Heart failure c

80 180

Heart failure c Mild







From Reference 14. Reference 17. c From References 14,16. Abbreviations: BNP = B-type natriuretic peptide, CHF = congestive heart failure, pg = picograms bFrom


What Role in the ICU? Plasma BNP has been studied primarily in patients who present to emergency departments with possible heart failure. Few studies have been performed in ICU patients. One study of ICU patients with sepsis showed that plasma BNP levels were useful in identifying patients with cardiac dysfunction (18). However, it is unlikely that plasma BNP will replace more traditional methods of evaluating cardiac function in the ICU. Plasma BNP levels might prove useful for monitoring the effectiveness of treatment for heart failure in the ICU or to identify patients who develop fluid overload. Until further studies are conducted in the ICU, the plasma BNP assay will remain a tool for the emergency department.

Management Strategies The management of heart failure described here is meant for patients with advanced or decompensated heart failure, where the cardiac output is compromised (stage 3 in Figure 14.2). The approach here is specifically designed for ICU patients: it is based on invasive hemodynamic measurements rather than symptoms and uses only drugs that are given by continuous intravenous infusion (19,20,21). The hemodynamic drugs in this chapter are presented in detail in Chapter 16: the dose ranges and actions of each drug are shown in Table 14.2. TABLE 14.2 Drugs Used to Manage Acute, Decompensated Heart Failure in the ICU*


Dose Range

Principal Effect


3–15 µg/kg/min

Positive inotropic effect and systemic vasodilatation


1–3 µg/kg/min 3–10 µg/kg/min

Renal vasodilatation and natriuresis Positive inotropic effect and systemic vasodilatation

>10 µg/kg/min

Systemic vasoconstriction


50 µg/kg bolus, then 0.25–1 µg/kg/min

Positive inotropic effect, lusitropic effect, and systemic vasodilatation


1–50 µg/min >50 µg/min

Venous vasodilatation Arterial vasodilatation


0.3–2 µg/kg/min

Systemic vasodilatation

*Includes only drugs given by continuous intravenous infusion.

Left-Sided (Systolic) Heart Failure The management of decompensated left-sided heart failure is traditionally designed for a systolic-type heart failure, even though some cases may involve diastolic failure. The recommendations here are based on three P.264 measurements: the pulmonary capillary wedge pressure (PCWP), the cardiac output (CO), and the arterial blood pressure (BP). Decompensated heart failure is associated with a high PCWP and a low CO, but the BP can vary. The management strategies that follow are based on the condition of the blood pressure (i.e., high, normal, or low).

High Blood Pressure Decompensated heart failure with elevated blood pressure is a common scenario in the early period after cardiopulmonary bypass surgery (22). P.265

Profile: High PCWP/Low CO/High BP Treatment: Vasodilator therapy with nitroprusside or nitroglycerin. If the PCWP remains above 20 mm Hg, add diuretic therapy with furosemide. Vasodilators like nitroprusside and nitroglycerin augment cardiac output by reducing ventricular afterload. The overall effect is a decrease in arterial blood pressure, an increase in cardiac output, and a decrease in ventricular filling pressure (20). Nitroprusside is a more effective vasodilator than nitroglycerin, but drug safety is a concern. The major problem with nitroprusside is cyanide toxicity (23), which is more common than suspected (see Chapter 16) and is particularly prevalent following cardiopulmonary bypass surgery. Nitroprusside is also not advised in patients with ischemic heart disease because the drug can produce a “coronary steal syndrome” (4). Nitroglycerin is a safer alternative to nitroprusside. Low infusion rates (,50 mg/min) produce venous vasodilation (which can reduce cardiac output further), and dose rates in excess of 50 mg/min are usually required to produce effective arterial vasodilation. The major drawback with nitroglycerin infusions is the development of tolerance, which can appear after 16 to 24 hours of continuous drug administration (4). Vasodilator therapy with angiotensin-converting-enzyme (ACE) inhibitors, while beneficial in the long-term management of left heart failure, is not recommended for the acute management of decompensated left heart failure (4). Diuretic therapy with furosemide is indicated only if vasodilator therapy does not reduce the wedge pressure to the desired level. The desired wedge pressure in left heart failure is the highest pressure that will augment cardiac output without producing pulmonary edema. This is shown in Figure 14.6 as the highest point on the lower (heart failure) curve that does not enter the shaded (pulmonary edema) region. The desired or optimal wedge pressure in left heart failure is 18 to 20 mm Hg (24). Therefore diuretic therapy is indicated only if the wedge pressure during vasodilator therapy remains above 20 mm Hg. The features of diuretic therapy for decompensated heart failure are described later.

Normal Blood Pressure Decompensated heart failure with a normal blood pressure is the usual presentation of heart failure resulting from ischemic heart disease, acute myocarditis, and the advanced stages of chronic cardiomyopathy.

Figure 14.6 Ventricular function curves for the normal and failing left ventricle. Arrows show the expected changes associated with each type of drug therapy. The shaded area indicates the usual region where pulmonary edema becomes apparent.

View Figure

Profile: High PCWP/Low CO/Normal BP Treatment: Inodilator therapy with dobutamine or milrinone, or vasodilator therapy with nitroglycerin. If the PCWP does not decrease to 75 years, heart failure, new or worsening mitral regurgitation, markedly elevated cardiac troponin levels, and cardiogenic shock (3). The greatest benefits occur when these agents are used in conjunction with angioplasty (1,2,3,38). Abciximab is recommended only when angioplasty is planned and seems a favorite of cardiologists. In the catheterization lab, the initial bolus of abciximab is given after the arterial sheath is placed, and the abciximab infusion is continued for 12 hours after the procedure (39). Platelet glycoprotein inhibitors are gaining popularity in patients with ST-elevation MI (STEMI), and are usually given in combination with angioplasty or thrombolytic therapy (2,3,37). In the future, expect platelet glycoprotein inhibitors to be combined with low-dose fibrinolytic agents as a prelude to coronary angioplasty (so-called “facilitated angioplasty”).

Adverse Effects The major risk with platelet glycoprotein inhibitors is bleeding. The incidence of bleeding from these agents is difficult to assess because they are often used in combination with aspirin and heparin. Most of the bleeding is mucocutaneous, and intracranial hemorrhage is not a risk with these agents (7,38). Thrombocytopenia is reported in up to 2% of patients who receive abciximab and is more common with repeated use of the drug (38). Active bleeding is an absolute contraindication to platelet glycoprotein inhibitors. Relative contraindications include major surgery within the past 3 months, stroke in the past 6 months, systolic blood pressure >180 mm Hg or diastolic pressure >110 mm Hg, and severe thrombocytopenia (38).

Early Complications The appearance of decompensated heart failure and cardiogenic shock in the first few days after an acute MI is an ominous sign and usually indicates a mechanical problem like acute mitral regurgitation or cardiac pump failure. Echocardiography is usually needed to uncover the problem, but the mortality in these conditions is high despite timely interventions. P.335

Mechanical Complications

Mechanical complications are usually the result of transmural (ST-elevation) MI. All are serious, and all require prompt action. Acute mitral regurgitation is the result of papillary muscle rupture and presents with the sudden onset of pulmonary edema and the characteristic holosystolic murmur radiating to the axilla. The pulmonary artery occlusion pressure should show prominent V waves, but this can be a non-specific finding. Diagnosis is by echocardiography, and arterial vasodilators (e.g., hydralazine) are used to relieve pulmonary edema pending surgery. Mortality is 70% without surgery and 40% with surgery (39). Ventricular septal rupture can occur anytime in the first 5 days after acute MI. The diagnosis can be elusive without cardiac ultrasound. There is a step-up in O 2 saturation from right atrial to pulmonary artery blood, but this is rarely measured. Initial management involves vasodilator (e.g., nitroglycerin) infusions and the intraaortic balloon pump if needed. Mortality is 90% without surgery and 20% to 50% with surgery (2). Ventricular free wall rupture occurs in up to 6% of cases of STEMI and is more common with anterior MI, fibrinolytic or steroid therapy, and advanced age (2). The first signs of trouble are usually return of chest pains and new ST-segment abnormalities on the ECG. Accumulation of blood in the pericardium often leads to rapid deterioration and cardiovascular collapse from pericardial tamponade. Diagnosis is made by cardiac ultrasound (if time permits), and prompt pericardiocentesis combined with aggressive volume resuscitation is required for hemodynamic support. Immediate surgery is the only course of action, but fewer than half of the patients survive despite surgery (2).

Pump Failure About 15% of cases of acute MI result in cardiac pump failure and cardiogenic shock (40). Management involves hemodynamic support (usually with intraaortic balloon counterpulsation) followed by reperfusion using coronary angioplasty or coronary bypass surgery. Despite the best intentions, the mortality in this situation is 60 to 80% (40).

Hemodynamic Support Hemodynamic support should be designed to augment cardiac output without increasing myocardial oxygen consumption. Table 17.6 shows the effects of hemodynamic support on the determinants of myocardial O2 consumption (preload, contractility, afterload, and heart rate) in decompensated heart failure and cardiogenic shock. As judged by the net effect on myocardial O2 consumption, vasodilator therapy is superior to dobutamine in heart failure, and the intra-aortic balloon pump (IABP) is superior to dopamine in cardiogenic shock. (See Chapter 14 for more information on the treatment of cardiac pump failure.) TABLE 17.6 Hemodynamic Support and Myocardial O2 Consumption


Heart Failure



Cardiogenic Shock IABP















Heart rate



Net effect on myocardial VO 2





IABP = Intra-aortic balloon pump, VO2 5 oxygen consumption.


Emergency Revascularization The ACC/AHA guidelines recommend coronary angioplasty when cardiogenic shock appears within 36 hours of acute MI and when the angioplasty can be performed within 18 hours of the onset of shock (2). Coronary artery bypass surgery is considered if the cardiac catheterization reveals multivessel disease that is not amenable to angioplasty or disease involving the left main coronary artery (2).

Arrhythmias Disturbances of cardiac rhythm are common after acute MI and are not suppressed by prophylactic use of lidocaine (2). The management of serious arrhythmias is described in the next chapter.

Acute Aortic Dissection Aortic dissection is included in this chapter because the clinical presentation can be mistaken as an acute coronary syndrome, and the condition is often fatal if missed.

Clinical Presentation The most common complaint is abrupt onset of chest pain. The pain is often sharp and is described as “ripping or tearing” (mimicking the underlying process) in about 50% of cases (41). Radiation to the jaws and arms is uncommon. The pain can subside spontaneously for hours to days (41,42), and this can be a source of missed diagnoses. The return of the pain after a pain-free interval is often a sign of impending aortic rupture.

Clinical Findings Hypertension and aortic insufficiency are each present in about 50% of cases, and hypotension is reported in 25% of cases (41,42). Dissection can cause obstruction of the left subclavian artery, leading to blood pressure P.337 differences in the arms, but this finding can be absent in up to 85% of cases (42). Obstruction involving other arteries in the chest can lead to stroke and coronary insufficiency.

Diagnosis Mediastinal widening on chest x-ray (present in 60% of cases) often raises suspicion for dissection (42). However, the diagnosis requires one of four imaging modalities ( 43): magnetic resonance imaging (MRI) (sensitivity and specificity, 98%), transesophageal echocardiography (sensitivity, 98%; specificity, 77%), contrast-enhanced computed tomography (sensitivity, 94%; specificity, 87%), and aortography (sensitivity, 88%; specificity, 94%). Thus MRI is the diagnostic modality of choice for aortic dissection, but the immediate availability of MRI is limited in some hospitals, and helical CT and transesophageal ultrasound are high-yield alternatives to MRI. Aortography is the least sensitive but provides valuable information for the operating surgeon.

Management Acute dissection in the ascending aorta is a surgical emergency. Prompt control of hypertension is advantageous prior to surgery to reduce the risk of aortic rupture. Increased flow rates in the aorta create shear forces that promote further dissection, so blood pressure reduction should not be accompanied by increased cardiac output. This can be accomplished with the drug regimens shown in Table 17.7 (41,42). One regimen uses a vasodilator (nitroprusside) infusion combined with a ß-blocker (esmolol) infusion. The ß-blocker is given first to block the vasodilator-induced increase in cardiac output. Esmolol is used as the ß-blocker because it has a short duration of action (9 minutes) and is easy to titrate. It can also be stopped just prior to surgery without the risk of residual cardiac suppression during surgery. Single-drug therapy with a combined a and ß-receptor antagonist (labetalol) is also effective and is easier to use than the combination drug regimen. TABLE 17.7 Treating Hypertension in Aortic Dissection

Combined therapy with ß-blocker and vasodilator: Start with esmolol:

500 µ/kg IV bolus and follow with 50 µg/kg/min. Increase infusion by 25 µg/kg/min every 5 min until heart rate 60–80 bpm. Maximum dose rate is 200 µg/kg/min.

Add nitroprusside:

Start infusion at 0.2 µg/kg/min and titrate upward to desired effect. See the nitroprusside dosage chart in Table 16.6.

Monotherapy with combined a-ß receptor antagonist: Labetalol:

20 mg IV over 2 min, then infuse 1–2 mg/min to desired effect and stop infusion. Maximum cumulative dose is 300 mg.


A Final Word The discovery that acute myocardial infarction is the result of blood clots that obstruct coronary arteries has one important implication (besides the improved therapeutic approach to myocardial infarction) that seems overlooked. It disputes the traditional teaching that myocardial infarction is the result of a generalized imbalance between myocardial O2 delivery and O2 consumption. This distinction is important because the O2-imbalance paradigm is the basis for the overzealous use of oxygen breathing and blood transfusions in patients with coronary artery disease. Blood clots (from ruptured atherosclerotic plaques) cause heart attacks, not hypoxia or anemia. If you have ever wondered why heart attacks are uncommon in patients with progressive shock and multiorgan failure, you now have the answer.

References Clinical Practice Guidelines 1. 2005 American Heart Association Guidelines for Cardiopulmonary Resuscitation and Emergency Cardiovascular Care. Part 8: Stabilization of the patient with acute coronary syndromes. Circulation 2005;112(suppl I):IV89–IV110.

(Available online @ 2. Antman EM, Anbe DT, Armstrong PW, et al. ACC/AHA guidelines for the management of patients with ST-elevation myocardial infarction: executive summary: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Writing Committee to Revise the 1999 Guidelines for the Management of Patients with Acute Myocardial Infarction). Circulation 2004;110:588–636. (Available at Full TextBibliographic Links 3. Braunwald E, Antman EM, Beasley JW, et al. ACC/AHA 2002 guideline update for the management of patients with unstable angina and non-ST-segment myocardial infarction: summary article: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines (Committee on the Management of Patients with Unstable Angina). J Am Coll Cardiol 2002;40:1366–1374. Full TextBibliographic Links

Acute Coronary Syndromes 4. Davies MJ, Thomas AC. Plaque fissuring: the cause of acute myocardial infarction. Br Heart J 1985;53:363–373. Bibliographic Links 5. Van der Wal AC, Becker AE, van der Loos CM, et al. Site of intimal rupture or erosion of thrombosed coronary atherosclerotic plaques is characterized by an inflammatory process irrespective of the dominant plaque morphology. Circulation 1994;89:36–44. Ovid Full TextBibliographic Links 6. Malek AM, Alper SL, Izumo S. Hemodynamic shear stress and its role in atherosclerosis. JAMA 1999;282:2035–2042. Ovid Full TextBibliographic Links P.339

Routine Measures 7. Patrono C, Coller B, Fitzgerald G, et al. Platelet-active drugs: the relationships among dose, effectiveness, and side effects. Chest 2004;126:234S–264S. Full TextBibliographic Links 8. ISIS-2 (Second International Study of Infarct Survival) Collaborative group. Randomized trial of intravenous streptokinase, oral aspirin, both, or neither among 17,187 cases os suspected acute myocardial infarction: ISIS-2. Lancet

1988;2:349–360. Full TextBibliographic Links 9. Roux S, Christellar S, Ludin E. Effects of aspirin on coronary reocclusion and recurrent ischemia after thrombolysis: a meta-analysis. J Am Coll Cardiol 1992;19:671–677. Bibliographic Links 10. Clopidogrel in Unstable Angina to Prevent Recurrent Events Trial Investigators. Effects of clopidogrel in addition to aspirin in patients with acute coronary syndromes without ST-segment elevation. N Engl J Med 2001;345:494–502. Ovid Full TextBibliographic Links 11. Kloner RA, Hale S. Unraveling the complex effects of cocaine on the heart. Circulation 1993;87:1046–1047. Ovid Full TextBibliographic Links 12. AHFS. Metoprolol succinate and metoprolol tartrate. In McEvoy GK, ed. AHFS drug information, 2001. Bethesda, MD: American Society for Health System Pharmacists, 2001:1622–1629.

13. ACE Inhibitor Myocardial Infarction Collaborative Group. Indications for ACE inhibitors in the early treatment of acute myocardial infarction: systematic overview of individual data from 100,000 patients in randomized trials. Circulation 1998;97:2202–2212. Ovid Full TextBibliographic Links 14. Gruppo Italiano per lo Studio della Sopravvivenza nell'infarto Miocardico (GISSI). GISSI-3. Effects of lisinopril and transdermal glyceryl trinitrate singly and together on 6-week mortality and ventricular function after acute myocardial infarction. Lancet 1994;343:1115–1122. Full TextBibliographic Links 15. Pfeffer MA, McMurray JJ, Velazquez EJ, et al., for the Valsartan in Acute Myocardial Infarction Trial Investigators. Valsartan, captopril, or both in myocardial infarction complicated by heart failure, left ventricular dysfunction, or both. N Engl J Med 2003;349:1893–1906. Ovid Full TextBibliographic Links

Thrombolytic Therapy 16. Anderson HV, Willerson JT. Thrombolysis in acute myocardial infarction. N Engl J Med 1993;329:703–725. Ovid Full TextBibliographic Links

17. Gruppo Italiano per lo Studio della Streptochinasi nell'Infarto Miocardico (GISSI). Effectiveness of intravenous thrombolytic treatment in acute myocardial infarction. Lancet 1986;1:397–401. Bibliographic Links 18. Fibrinolytic Therapy Trialists Collaborative Group. Indications for fibrinolytic therapy in suspected acute myocardial infarction: collaborative overview of early mortality and major morbidity results from all randomized trials of more than 1000 patients. Lancet 1994;343:311–322. Full TextBibliographic Links 19. Boden WE, Kleiger RE, Gibson RS, et al. Electrocardiographic evolution of posterior acute myocardial infarction: importance of early precordial ST-segment depression. Am J Cardiol 1987;59:782–787. Bibliographic Links 20. Guidry JR, Raschke R, Morkunas AR. Anticoagulants and thrombolytics. In: Blumer JL, Bond GR, eds. Toxic effects of drugs in the ICU: critical care clinics. Vol. 7. Philadelphia: WB Saunders, 1991:533–554. P.340 21. GUSTO Investigators. An international randomized trial comparing four thrombolytic strategies for acute myocardial infarction. N Engl J Med 1993;329:673–682. Ovid Full TextBibliographic Links 22. Llevadot J, Giugliano RP, Antman EM. Bolus fibrinolytic therapy in acute myocardial infarction. JAMA 2001;286:442–449. Ovid Full TextBibliographic Links 23. Smalling RW, Bode C, Kalbfleisch J, et al. More rapid, complete, and stable coronary thrombolysis with bolus administration of reteplase compared with alteplase infusion in acute myocardial infarction. Circulation 1995;91:2725–2732. Ovid Full TextBibliographic Links 24. GUSTO-III Investigators. An international, multicenter, randomized comparison of reteplase with alteplase for acute myocardial infarction. N Engl J Med 1997;337:1118–1123. Bibliographic Links 25. Assessment of the Safety and Efficacy of a New Thrombolytic (ASSENT-2) Investigators. Single-bolus tenecteplase compared with front-loaded alteplase in acute myocardial infarction. Lancet 1999;354:716–722. Full TextBibliographic Links 26. Young GP, Hoffman JR. Thrombolytic therapy. Emerg Med Clin


27. Topol EJ. Acute myocardial infarction: thrombolysis. Heart 2000;83:122–126. Bibliographic Links

Coronary Angioplasty 28. The GUSTO IIb Angioplasty Substudy Investigators. A clinical trial comparing primary coronary angioplasty with tissue plasminogen activator for acute myocardial infarction. N Engl J Med 1997;336:1621–1628. Ovid Full TextBibliographic Links 29. Keeley EC, Boura JA, Grines CL. Primary angioplasty versus intravenous thrombolytic therapy for acute myocardial infarction: a quantitative review of 23 randomized trials. Lancet 2003;361:13–20. Full TextBibliographic Links 30. Stone GW, Cox D, Garcia E, et al. Normal flow (TIMI-3) before mechanical reperfusion therapy is an independent determinant of survival in acute myocardial infarction. Circulation 2001;104:636–641. Ovid Full TextBibliographic Links 31. Cannon CP, Gibson CM, Lambrew CT, et al. Relationship of symptom onset to balloon time and door-to-balloon time with mortality in patients undergoing angioplasty for acute myocardial infarction. JAMA 2000;283:2941–2947. Ovid Full TextBibliographic Links 32. Meyer MC. Reperfusion strategies for ST-segment elevation myocardial infarction: an overview of current therapeutic options. Part II: Mechanical reperfusion. Emergency Medicine Reports, Vol 25, April 2004 (Accessed on 1/2/2005 at 33. Andersen HR, Nielsen TT, Rasmussen K, et al. for the DANAMI-2 Investi-gators. A comparison of coronary angioplasty with fibrinolytic therapy in acute myocardial infarction. N Engl J Med 2003;349:733–742. Ovid Full TextBibliographic Links

Adjuncts to Reperfusion Therapy 34. Wong GC, Gugliano RP, Antman EM. Use of low-molecular-weight heparins in the management of acute coronary syndromes and percutaneous coronary intervention. JAMA 2003;289:331–342.

Ovid Full TextBibliographic Links P.341 35. Petersen JL, Mahaffey KW, Hasselblad V, et al. Efficacy and bleeding complications among patients randomized to enoxaparin or unfractionated heparin for antithrombin therapy in non-ST-segment elevation acute coronary syndromes. JAMA 2004;292:89–96. Ovid Full TextBibliographic Links 36. Theroux P, Welsh RC. Meta-analysis of randomized trials comparing enoxaparin versus unfractionated heparin as adjunctive therapy to fibrinolysis in ST-elevation acute myocardial infarction. Am J Cardiol 2003;91:860–864. Bibliographic Links 37. Assessment of the Safety and Efficacy of a New Thrombolytic Regimen (ASSENT-3) Investigators. Efficacy and safety of tenectaplase in combination with enoxaparin, abciximab, or unfractionated heparin: the ASSENT-3 randomized trial in acute myocardial infarction. Lancet 2001;358:605–613. Full TextBibliographic Links 38. Bhatt DL, Topol EJ. Current role of platelet glycoprotein IIb/IIIa inhibitors in acute coronary syndromes. JAMA 2000;284:1549–1558. Ovid Full TextBibliographic Links

Early Complications 39. Thompson CR, Buller CE, Sleeper LA, et al. Cardiogenic shock due to acute severe mitral regurgitation complicating acute myocardial infarction: a report from the SHOCK trial registry. J Am Coll Cardiol 2000;36:1104–1109. Full TextBibliographic Links 40. Samuels LF, Darze ES. Management of acute cardiogenic shock. Cardiol Clin 2003;21:43–49. Bibliographic Links

Aortic Dissection 41. Khan IA, Nair CK. Clinical, diagnostic, and management perspectives of aortic dissection. Chest 2002;122:311–328. Full TextBibliographic Links 42. Knaut AL, Cleveland JC. Aortic emergencies. Emerg Med Clin N Am 2003; 21:817–845.

43. Zegel HG, Chmielewski S, Freiman DB. The imaging evaluation of thoracic aortic dissection. Appl Radiol 1995;(June):15–25. Bibliographic Links

Chapter 18 Tachyarrhythmias Acute arrhythmias are the gremlins of the ICU because they pop up unexpectedly, create some havoc, and are often gone in a flash. The arrhythmias that create the most havoc are the ones that produce rapid heart rates: the tachyarrhythmias. This chapter describes the acute management of tachyarrhythmias using clinical practice guidelines developed by consensus groups in the United States and Europe. The published guidelines are listed in the bibliography at the end of chapter (1,2,3), along with Internet addresses where they can be downloaded at no cost.

Classification Tachycardias (heart rate above 100 beats/minute) can be the result of increased automaticity in pacemaker cells (e.g., sinus tachycardias), triggered activity (e.g., ectopic impulses), or a process known as re-entry, where a triggered impulse encounters a pathway that blocks propagation in the forward direction but allows the impulse to pass in the return (retrograde) direction. Such retrograde transmission allows a triggered impulse to propagate continually, creating a self-sustaining tachycardia. Re-entry is the most common cause of clinically significant tachycardias. Tachycardias are classified according to the site of impulse generation in relation to the atrioventricular (AV) conduction system. Supraventricular tachycardias (SVT) originate above the AV conduction system and have a normal QRS duration (= 0.12 seconds), while ventricular tachycardias (VT) originate below the AV conduction system and have a prolonged QRS duration (. 0.12 seconds). Each type of tachycardia can then be subdivided according to regularity of the rhythm (i.e., the regularity of the R–R interval on the ECG). The classification of tachycardias based on the QRS duration and the regularity of the R-R interval is shown in Figure 18.1.

Narrow QRS Complex Tachycardia The tachycardias associated with a QRS duration of 0.12 seconds or less include sinus tachycardia, atrial tachycardia, AV nodal re-entrant tachycardia (also called paroxysmal supraventricular tachycardia), atrial flutter, and atrial fibrillation.

Figure 18.1 Classification of the tachycardias based on the QRS duration and the regularity of the R–R interval on the electrocardiogram.

View Figure


Regular Rhythm If the R-R intervals are uniform in length (indicating a regular rhythm), the possible arrhythmias include sinus tachycardia, AV nodal re-entrant tachycardia, or atrial flutter with a fixed (2:1, 3:1) AV block. The atrial activity can help to identify each of these rhythms. Uniform P waves with a fixed P–R interval is characteristic of sinus tachycardia; the absence of P waves suggests an AV nodal re-entrant tachycardia (see Figure 18.2), and sawtooth waves are characteristic of atrial flutter. If the R-R intervals are not uniform in length (indicating an irregular rhythm), the most likely arrhythmias are multifocal atrial tachycardia P.345 and atrial fibrillation. Once again, the atrial activity helps to identify each of these rhythms. Multifocal atrial tachycardia has multiple P wave morphologies and nonuniform PR intervals (Panel A, Figure 18.3), while atrial fibrillation has fibrillatory atrial waves and no identifiable P waves (Panel B, Figure 18.3). The rhythm in atrial fibrillation is highly irregular and is sometimes described as an “irregularly irregular” rhythm (which indicates that no two R-R intervals have the same length).

View Figure

Figure 18.2 AV nodal re-entrant tachycardia, which is also called a paroxysmal supraventricular tachycardia. Note the absence of P waves, which are hidden in the QRS complexes.

Figure 18.3 Multifocal atrial tachycardia (panel A) and atrial fibrillation (panel B). (From the CD ROM that accompanies Critical care nursing: A holistic approach. Philadelphia: Lippincott Williams & Wilkins, 2005.)

View Figure

Wide QRS Complex Tachycardia A tachycardia with a QRS duration > 0.12 seconds is either ventricular tachycardia (VT) or SVT with aberrant (prolonged) AV conduction. VT is characterized by a regular rhythm and the presence of AV dissociation, while SVT with aberrant conduction can have a regular or irregular rhythm (depending on the rhythm of the inciting SVT). These two arrhythmias can look remarkably similar, as presented later in the chapter.

Sinus Tachycardia Increased automaticity in the pacemaker cells of the sinoatrial node produces a regular, narrow-complex tachycardia with a gradual onset and rate of 100 to 140 beats/minute. The ECG shows uniform P waves and a fixed P–R interval. Sinus tachycardia can also be the result of re-entry into the sinus node. This variant sinus tachycardia has an abrupt onset, P.346 but is otherwise indistinguishable from the increased automaticity type of sinus tachycardia.

Management Sinus tachycardia is often a response to a systemic illness. It is usually well tolerated (cardiac filling is usually not compromised until the heart rate rises above 180 beats/minute) (4) and does not require primary treatment. The primary goal of management is to identify and treat the associated illness. Possible sources of sinus tachycardia in the ICU include systemic infection and inflammation, hypovolemia, and adrenergic drugs. The major indication for slowing a sinus tachycardia is the presence of myocardial ischemia or infarction. In this situation, ß-receptor antagonists can be used to slow the heart rate. (See Chapter 17, Figure 17.2, for an effective ß-blocker regimen in acute coronary syndromes.) Because these agents also depress ventricular function, they are not recommended for sinus tachycardia associated with systolic heart failure.

Atrial Fibrillation Atrial fibrillation (AF) is the most common cardiac arrhythmia in the general population. (Atrial flutter is considered here as a more organized form of atrial fibrillation rather than a distinct arrhythmia). An estimated 2.2 million adults (or 1% of the adult population) have AF (3). Most are elderly (median age, 75 years) and have either ischemic heart disease, valvular disease, or cardiomyopathy. Contrary to popular perception, few have hyperactive thyroid disease (5). About 15% of patients with AF are relatively young (less than 60 years of age) and have no predisposing conditions (3): this condition is known as lone atrial fibrillation.

Postoperative AF Postoperative AF is reported in 30 to 40% of patients undergoing coronary artery bypass surgery, and 60% of patients undergoing valve surgery, and usually appears in the first 4 postoperative days (6). The etiology is unclear, but risk factors include valvular surgery, advanced age, and failure to resume ß-blocker therapy after surgery. ß-blockers are preferred for rate control of AF in this setting (7). This arrhythmia is usually self-limited, and more than 90% of patients will convert to sinus rhythm within 6 to 8 weeks (3).

Adverse Consequences Contraction of the atria is responsible for 25% of the ventricular end-diastolic volume (preload) in the normal heart (4). This atrial contribution to ventricular filling is lost in AF. There is little consequence in the normal heart, but cardiac output can be impaired in patients with diastolic dysfunction due to a noncompliant or stiff ventricle (where ventricular P.347 filling volumes are already reduced). This effect is pronounced at rapid heart rates (because of the reduced time for ventricular filling). The other notable complication of AF is thrombus formation in the left atrium, which can embolize to the cerebral circulation and produce an ischemic stroke. Atrial thrombosis can be demonstrated in 15% of patients who have AF for longer than 3 days (8), and about 6% of patients with chronic AF and certain risk factors (see later) suffer an embolic stroke each year without adequate anticoagulation (3,8). The indications for anticoagulation in AF are presented later.

Management Strategies The acute management of AF involves 3 strategies: (1) cardioversion to terminate the arrhythmia and restore normal sinus rhythm, (2) drug-induced reduction of the ventricular rate, and (3) anticoagulation to prevent thromboembolism. The following presentation is organized according to these strategies.


Cardioversion can be accomplished by applying electric shocks (electrical cardioversion) or administering an antiarrhythmic agent (pharmacological cardioversion).

Electrical Cardioversion Immediate cardioversion using direct-current (DC) electric shocks is indicated for cases of AF associated with severe hemodynamic compromise (hypotension or decompensated heart failure). This procedure is both painful and anxiety-provoking and, if tolerated, pre-medication with a benzodiazepine (e.g., midazolam) and/or an opiate (morphine or fentanyl) is indicated. The individual shocks should be synchronized to the R wave of the QRS complex to prevent electrical stimulation during the vulnerable period of ventricular repolarization, which usually coincides with the peak of the T wave (3). The following protocols are recommended (3). 1. For monophasic shocks, begin with 200 joules (J) for atrial fibrillation and 50 J for atrial flutter. If additional shocks are needed, increase the energy level of each successive shock by 100 J until a maximum shock strength of 400 J is reached. Wait at least one minute between shocks to minimize the risk of cardiac ischemia. 2. For biphasic shocks (which are the waveforms used in many of the newer defibrillator machines), use only half the energy recommended for monophasic shocks. These regimens should result in successful cardioversion in about 90% of cases ( 3). P.348

Pharmacologic Cardioversion Acute pharmacologic cardioversion may be appropriate for first episodes of AF that are less than 48 hours in duration and are not associated with hemodynamic compromise or evidence of cardiac ischemia. In this situation, conversion to sinus rhythm will avoid the need for anticoagulation (see later) and can prevent atrial remodeling that predisposes to recurrent AF (9). However, over 50% of cases of recent-onset AF convert spontaneously to sinus rhythm in the first 72 hours (10), so cardioversion is not necessary in most cases of recent-onset AF unless the symptoms are distressing. Several antiarrhythmic agents are recommended for the acute termination of AF (flecainide, propafenone, ibutilide, difetilide, and amiodarone), but the only one with a success rate higher than 5 or 10% is ibutilide. When patients with recent-onset AF are given ibutilide in the dosing regimen shown in Table 18.1, over 50% will convert to sinus rhythm, and 80% will show a response within 30 minutes of drug administration (11). The only risk associated with ibutilide is torsade de pointes (described later), which is reported in 4% of cases (11). (For information P.349 on the other antiarrhythmic agents recommended for cardioversion of AF, see Reference #3.) TABLE 18.1 Intravenous Drug Regimens for Acute Management of Atrial Fibrillation†


Dose Regimen



1 mg IV over 10 min and repeat once if needed.

The best agent available for acute cardioversion of AF. Torsade de pointes reported in 4% of cases


0.25 mg/kg IV over 2 min, then 0.35 mg/kg 15 minutes later if needed. Follow with infusion of 5–15 mg/hr for 24 hrs.

Effective rate control in > 95% of patients. Has negative inotropic actions, but can be used safely in patients with heart failure.


500 µg/kg IV over 1 min, then infuse at 50 µg/kg/min. Increase dose rate by 25 µg/kg/min every 5 min if needed to maximum of 200 µg/kg/min.

Ultra-short-acting ß-blocker that permits rapid dose titration to desired effect.


2.5 to 5 mg IV over 2 min. Repeat every 10–15 min if needed to total of 3 doses.

Easy to use, but bolus dosing is not optimal for exact rate control.


300 mg IV over 15 min, then 45 mg/hr for 24 hrs. ‡

A suitable alternative for patients who do not tolerate more effective rate-reducing drugs.

From the recommendations in Reference 3.


from Reference 17.

Intravenous amiodarone is also recommended for acute termination of AF despite evidence of variable and limited efficacy. Bolus administration of amiodarone results in acute cardioversion of AF in less than 5% of cases (12). This agent may be more useful

for acute rate control in AF, as described later.

Controlling the Heart Rate The acute management of AF (particularly chronic or recurrent AF) is most often aimed at reducing the ventricular rate into the range of 60 to 80 bpm (3). If an arterial catheter is in place, monitoring the systolic blood pressure can provide a more physiological end-point for rate control. The systolic blood pressure is a reflection of the stroke volume, and the principal determinant of stroke volume is ventricular end-diastolic volume (this is the Frank-Starling relationship described in Chapter 1). When the heart rate in AF is slow enough to allow for adequate ventricular filling during each period of diastole, the systolic blood pressure (stroke volume) should remain constant with each heart beat. Therefore a constant systolic blood pressure with each heart beat can be used as an end-point of rate control in AF. The drugs that are used for acute rate control in AF are shown in Table 18.1. These drugs are either calcium-channel blockers (diltiazem) or ß-blockers (esmolol and metoprolol), and they act by prolonging conduction through the atrioventricular node, which slows the ventricular response to the rapid atrial rate.

Calcium-Channel Blockers Verapamil was the original calcium-channel blocker used for acute rate control in AF, but diltiazem (Cardizem) is now preferred because it produces less myocardial depression and is less likely to produce hypotension (3). When given in appropriate doses, diltiazem will produce satisfactory rate control in 85% of patients with AF (13). The response to a bolus dose of diltiazem is evident within 5 minutes (see Figure 18.4), and the effect dissipates over the next 1 – 3 hours (14). Because the response is transient, the initial bolus dose of diltiazem should be followed by a continuous infusion. Although diltiazem has mild negative inotropic actions, it has been used safely in patients with heart failure (15).

ß-Receptor Antagonists ß-blockers are the preferred agents for rate control when AF is associated with hyperadrenergic states (such as acute MI and post-cardiac surgery) ( 3,7). Two ß-blockers with proven efficacy in AF are esmolol (Brevibloc) and metoprolol (Lopressor), and their dosing regimens are shown in Table 18.1. Both are cardioselective agents that preferentially block ß-1 receptors in the heart. Esmolol is the preferred agent for acute rate control because it is ultra–short-acting (serum half-life of 9 minutes), and the infusion rate can be titrated rapidly to maintain the target heart P.350 rate (16). Because ß-blockers and calcium-channel-blockers both have cardiodepressant effects, combined therapy with ß-blockers and calcium-channel blockers should be avoided.

Figure 18.4 Comparative effects of IV diltiazem (same dose as in Table 18.1) and IV digoxin (0.5 mg in two divided doses over 6 hours) on acute control of heart rate in patients with recent-onset AF. (Adapted from data in Schreck DM, Rivera AR, Tricarico VJ, et al. Emergency treatment of atrial fibrillation and flutter: comparison of IV digoxin versus IV diltiazem. Ann Emerg Med 1995;25:127.)

View Figure

Amiodarone Amiodarone can prolong conduction in the AV node and reduce the ventricular rate in patients with AF. When given in the dosing regimen shown in Table 18.1, intravenous amiodarone can produce an acute reduction in heart rate in 75% of patients with recent-onset AF (17). Although not as effective as diltiazem, amiodarone produces less cardiac depression and is less likely to produce hypotension than diltiazem (17). As a result, amiodarone may be a suitable alternative when other rate-controlling drugs are not tolerated. The potential side effects of short-term intravenous amiodarone include hypotension (15%), infusion phlebitis (15%), bradycardia (5%), and elevated liver enzymes (3%) (11,18). Hypotension is the most common side effect and is related to the vasodilator actions of amiodarone and the solvent (polysorbate 80 surfactant) used to enhance water solubility of the injectable drug (18). Hypotension can usually be managed by P.351 decreasing the infusion rate or briefly stopping the infusion. The other common side effect is infusion phlebitis, which can be prevented by infusing amiodarone through a large, central vein. Amiodarone also has several drug interactions (18). It increases the serum concentrations of digoxin, warfarin, fentanyl, quinidine, procainamide, and cyclosporine. Many of the interactions are the result of amiodarone's metabolism via the cytochrome P450 enzyme system in the liver. The digoxin and warfarin interactions are the most important in the ICU.

Digoxin Digoxin, by virtue of its ability to prolong AV conduction, has been a popular and effective agent for long-term rate control in AF. However, it should not be used for immediate rate control in AF because of its delayed onset of action. This is demonstrated in Figure 18.4, which shows the results of a study comparing the effects of IV diltiazem and IV digoxin on acute rate control in patients with recent-onset AF. The heart rate decreased to below 100 bpm (the target rate) promptly in the patients who received diltiazem, while the heart rate remained above 110 bpm after 6 hours in the patients who received digoxin. These results demonstrate that digoxin is ineffective for acute rate control in AF.

Anticoagulation Cerebral embolic stroke (mentioned earlier) is the most devastating complication of AF. Each year about 6% of patients with AF will suffer an embolic stroke if they have certain high-risk conditions for thromboembolism. This can be reduced by 3% with warfarin anticoagulation to achieve an INR between 2.0 to 3.0 (19). In other words, therapeutic anticoagulation with warfarin in high-risk patients is associated with 3 fewer strokes for every 100 patients treated. This benefit requires strict monitoring of warfarin anticoagulation to keep the INR in the therapeutic range (2.0 – 3.0).

Risk Stratification Table 18.2 shows the different antithrombotic strategies based on the risk factors for thomboembolism in patients with AF (3). The patients with the highest risk for thromboembolism (annual stroke rate > 6%) that will benefit from warfarin anticoagulation include those with rheumatic valvular disease, prosthetic valves, a prior history of thromboembolism, heart failure with comorbid conditions, or advanced age (. 75 years of age). The conditions with a low risk of thromboembolism (annual stroke rate < 2%) can be treated with daily aspirin. The risk of thromboembolism is lowest (the same as the general population) in patients with AF who are younger than 60 years of age and have no evidence of heart disease. This condition is known as lone atrial fibrillation, and it does not require any form of antithrombotic therapy. The benefits of anticoagulation must always be weighed against the risk of hemorrhage, particularly intracerebral bleeding. Warfarin P.352 anticoagulation increases the yearly rate of intracerebral hemorrhage by < 1% (19), so the risk:benefit ratio favors anticoagulation in high-risk patients with AF. For patients with a predisposition to bleeding, the decision to anticoagulate is made on a case-by-case basis. TABLE 18.2 Risk-Based Antithrombotic Strategies for Patients with Atrial Fibrillation†

I. Oral Anticoagulation (INR = 2.0 – 3.0) Age > 75 years Age = 60 years plus diabetes or coronary artery disease Heart failure with LV ejection fraction, 48 hours, the risk of embolization with cardioversion is about 6%, and anticoagulation is recommended for 3 weeks before elective cardioversion.

Wolff–Parkinson–White Syndrome The WPW syndrome (short P–R interval and ? waves before the QRS) is characterized by recurrent supraventricular tachycardias that originate from an accessory (re-entrant) pathway in the AV conduction system (2). (Re-entrant tachycardias are described later in the chapter.) One of the tachycardias associated with WPW syndrome is atrial fibrillation. Agents that prolong AV conduction and produce effective rate control in conventional AF (such as the calcium-channel blockers) can paradoxically accelerate the ventricular rate (by blocking the wrong pathway) in P.353 patients with WPW syndrome. Thus in cases of AF associated with WPW syndrome, calcium-channel blockers and digoxin are contraindicated. The treatment of choice is electrical cardioversion or pharmacologic cardioversion with procainamide ( 20). The dosing regimen for procainamide is presented later.

Multifocal Atrial Tachycardia Multifocal atrial tachycardia (MAT) is characterized by multiple P wave morphologies and a variable P–R interval (see Figure 18.3). The ventricular rate is highly irregular, and MAT is easily confused with atrial fibrillation when atrial activity is not clearly displayed on the ECG. MAT is a disorder of the elderly (average age = 70), and over half of the cases occur in patients with chronic lung disease (21). The link with lung diseases may be partly due to the bronchodilator theophylline (22). Other associated conditions include magnesium and potassium depletion and coronary artery disease. (21).

Acute Management MAT can be a difficult arrhythmia to manage, but the steps listed below can be effective. 1. Discontinue theophylline (although this is no longer a popular bronchodilator). In one study, this maneuver resulted in conversion to sinus rhythm in half the patients with MAT (22). 2. Give intravenous magnesium (unless there is a contraindication) using the following protocol: 2 grams MgSO4 (in 50 mL saline) over 15 minutes, then 6 grams MgSO4 (in 500 mL saline) over 6 hours (23). In one study, this measure was effective in converting MAT to sinus rhythm in 88% of cases, even when serum magnesium levels were normal. The mechanism is unclear, but magnesium's actions as a calcium-channel blocker may be involved. 3. Correct hypomagnesemia and hypokalemia if they exist. If both disorders co-exist, the magnesium deficiency must be corrected before potassium replacement is started. (The reason for this is described in Chapter 34.) Use the following replacement protocol: infuse 2 mg MgSO 4 (in 50 mL saline) IV over 15 minutes, then infuse 40 mg potassium over 1 hour. 4. If the above measures are ineffective, give IV metoprolol if there is no evidence of COPD otherwise, give IV verapamil (a calcium-channel blocker). Metoprolol, given according to the regimen in Table 18.2, has been successful in converting MAT to sinus rhythm in 80% of cases (21). The verapamil dose is 75–150 µg/kg IV over 2 minutes (3). Verapamil converts MAT to sinus rhythm in less than 50% of cases, but it can also slow the ventricular rate. Watch for hypotension, which is a common side effect of verapamil. P.354

Paroxysmal Supraventricular Tachycardias Paroxysmal supraventricular tachycardias (PSVT) are narrow QRS complex tachycardias that are characterized by an abrupt onset and abrupt termination (unlike sinus tachycardia, which has a gradual onset and gradual resolution). These arrhythmias occur when there is an accessory pathway in the conduction system between atria and ventricles that conducts impulses at a different speed than the normal pathway. This difference in conduction velocities allows an impulse traveling down one pathway

(antegrade transmission) to travel up the other pathway (retrograde transmission). This circular transmission of impulses creates a rapid, self-sustaining, re-entrant tachycardia. The trigger is an ectopic atrial impulse that travels through either of the two pathways. There are 5 different types of PSVT, each characterized by the location of the accessory pathway. The most common PSVT is AV nodal re-entrant tachycardia, where the accessory pathway is located in the AV node.

AV Nodal Re-entrant Tachycardia AV nodal re-entrant tachycardia (AVNRT) is one of the most common rhythm disturbances in the general population. It most often occurs in subjects who have no evidence of structural heart disease and is more common in women (2). The onset is abrupt, and there may be distressing palpitations, but there is usually no evidence of heart failure or myocardial ischemia. The ECG shows a narrow QRS complex tachycardia with a regular rhythm and a rate between 140 and 220 bpm. There may be no evidence of atrial activity on the ECG (see Figure 18.2), which is a feature that distinguishes AVNRT from sinus tachycardia.

Acute Management Maneuvers that increase vagal tone can occasionally terminate AVNRT and can sometimes slow another type of tachycardia to reveal the diagnosis (e.g., sinus tachycardia). The vagal-enhancing maneuvers include the Valsalva maneuver (forced exhalation against a closed glottis), carotid massage, eyeball compression, and facial immersion in cold water (2). The value of these maneuvers is largely unproven, and some of the maneuvers (like eyeball compression or facial immersion in ice water) only add to the patient's distress and delay the termination of the arrhythmia. (Take a patient with AVNRT who is anxious and may be short of breath then stick the patient's face in a sink full of ice water and tell them to hold their breath, and you will know what I mean.) AVNRT can be terminated quickly by drugs that block the re-entrant pathway in the AV node. The most effective drugs are calcium-channel blockers (verapamil and diltiazem) and adenosine. These agents are equally effective for terminating AVNRT, but adenosine works faster and produces less cardiovascular depression than the calcium-channel blockers. TABLE 18.3 Intravenous Adenosine for Paroxysmal SVT

Indications: Termination of AV nodal re-entrant tachycardia, particularly in patients with Heart failure Hypotension Ongoing therapy with calcium channel blockers or ß-blockers WPW syndrome Contraindications: Asthma, AV block Dose: For delivery via peripheral veins: 1. Give 6 mg by rapid IV injection and flush with saline. 2. After 2 min, give a second dose of 12 mg if necessary. 3. The 12-mg dose can be repeated once. Dose adjustments: Decrease dose by 50% for: Injection into superior vena cava Patients receiving calcium blockers, ß-blockers, or dipyridamole Response: Onset of action, 0.5).

Increased CO2 Production An increase in CO2 production is usually related to oxidative metabolism, but non-metabolic CO 2 production is possible when extracellular acids generate hydrogen

ions that combine with bicarbonate ions and generate CO 2. Whatever the source, increased CO2 production is normally accompanied by an increase in minute ventilation, which eliminates the excess CO 2 and maintains a constant arterial PCO2. Therefore, excess CO2 production does not normally cause hypercapnia. However when CO 2 excretion is impaired (by neuromuscular weakness or lung disease), an increase in CO 2 production can lead to an increase in PaCO 2. Thus, increased CO 2 production is an important factor in promoting hypercapnia only in patients with a reduced ability to eliminate CO2.

Overfeeding Overfeeding, or the provision of calories in excess of daily needs, is a recognized cause of hypercapnia in patients with severe lung disease and acute respiratory failure (24). Nutrition-associated hypercapnia occurs predominantly in ventilator-dependent patients, and can delay weaning from mechanical ventilation. Overfeeding with carbohydrates is particularly problematic because oxidative metabolism of carbohydrates P.381 generates more carbon dioxide than the other nutrient substrates (lipids and proteins). Figure 19.7 Flow diagram for the evaluation of hypercapnia.

View Figure

Diagnostic Evaluation The bedside evaluation of hypercapnia is shown in Figure 19.7. The evaluation of hypercapnia, like hypoxemia, begins with the A-a PO2 gradient (25). A normal or unchanged A-a PO2 gradient indicates that the problem is alveolar hypoventilation (the same as described for the evaluation of hypoxemia). An increased A-a PO 2 gradient indicates a V/Q abnormality (an increase in dead space ventilation) that may or may not be accompanied by an increase in CO 2 production.

Measuring CO2 Production The rate of CO2 production (VCO2) can be measured at the bedside with specialized metabolic carts that are normally used to perform nutritional P.382 assessments. These carts are equipped with infrared devices that can measure the CO 2 in expired gas (much like the end-tidal CO 2 monitors described in Chapter 20), and can determine the volume of CO2 excreted per minute. In steady-state conditions, the rate of CO2 excretion is equivalent to the VCO2. The normal VCO2 is 90 to 130 L/minute/m2, which is roughly 80% of the VO2. As mentioned earlier, an increased VCO 2 is evidence for one of the following conditions: generalized hypermetabolism, overfeeding (excess calories), or organic acidoses.

A Final Word The measurement of arterial blood gases (PO2 and PCO2) enjoys a popularity that is undeserved, and this is particularly true for the arterial PO 2. It is important to remember that the arterial PO2 is not a useful measure for determining the amount of oxygen in the blood (this requires the hemoglobin concentration and the percent saturation of hemoglobin with oxygen, as described in Chapter 2). Instead, the PaO2 (along with the PaCO2) can be useful for evaluating gas exchange in the lungs. A more useful measurement for evaluating the oxygenation of blood is the pulse oximeter measurement of arterial oxyhemoglobin saturation (described in the next chapter).

References Introduction 1. Dale JC, Pruett SK. Phlebotomy: a minimalist approach. Mayo Clin Proc 1993;68:249–255. Bibliographic Links 2. Raffin TA. Indications for blood gas analysis. Ann Intern Med 1986;105:390–398. Full TextBibliographic Links

Pulmonary Gas Exchange 3. Dantzger DR. Pulmonary gas exchange. In: Dantzger DR, ed. Cardiopul-monary critical care. 2nd ed. Philadelphia: WB Saunders, 1991:25–43.

4. Lanken PN. Ventilation-perfusion relationships. In: Grippi MA, ed. Pulmon-ary pathophysiology. Philadelphia: JB Lippincott, 1995:195–210.

5. Buohuys A. Respiratory dead space. In: Fenn WO, Rahn H, eds. Handbook of physiology: respiration. Bethesda: American Physiological Society, 1964:699–714.

6. D'Alonzo GE, Dantzger DR. Mechanisms of abnormal gas exchange. Med Clin North Am 1983;67:557–571.

Quantitative Measures 7. Gammon RB, Jefferson LS. Interpretation of arterial oxygen tension. UpToDate Web Site, 2006. (Accessed 3/11/2006) P.383 8. Harris EA, Kenyon AM, Nisbet HD, et al. The normal alveolar-arterial oxygen tension gradient in man. Clin Sci 1974;46:89–104.

9. Gilbert R, Kreighley JF. The arterial/alveolar oxygen tension ratio: an index of gas exchange applicable to varying inspired oxygen concentrations. Am Rev Respir Dis 1974;109:142–145. Bibliographic Links 10. Carroll GC. Misapplication of the alveolar gas equation. N Engl J Med 1985;312:586. Bibliographic Links 11. Covelli HD, Nessan VJ, Tuttle WK. Oxygen derived variables in acute respiratory failure. Crit Care Med 1983;11:646–649. Bibliographic Links

Blood-Gas Variability 12. Hess D, Agarwal NN. Variability of blood gases, pulse oximeter saturation, and end-tidal carbon dioxide pressure in stable, mechanically ventilated trauma patients. J Clin Monit 1992;8:111–115. Bibliographic Links

13. Sasse SA, Chen P, Mahutte CK. Variability of arterial blood gas values over time in stable medical ICU patients. Chest 1994;106:187–193. Bibliographic Links

Hypoxemia 14. Duarte A, Bidani A. Evaluating hypoxemia in the critically ill. J Crit Illness 2005;20:91–93.

15. White AC. The evaluation and management of hypoxemia in the chronic critically ill patient. Clin Chest Med 2001;22:123–134. Bibliographic Links 16. Nowbar S, Burkhart KM, Gonzalez R, et al. Obesity-associated hypoventilation in hospitalized patients: prevalence, effects, and outcome. Am J Med 2004;116:1–7. Bibliographic Links 17. Rich MM, Raps EC, Bird SJ. Distinction between acute myopathy syndrome and critical illness polyneuropathy. Mayo Clin Proc 1995;70:198–199. Bibliographic Links 18. Bruschi C, Cerveri I, Zoia MC, et al. Reference values for maximum respiratory mouth pressures: a population-based study. Am Rev Respir Dis 1992; 146:790–793. Bibliographic Links 19. Baydur A. Respiratory muscle strength and control of ventilation in patients with neuromuscular disease. Chest 1991;99:330–338. Bibliographic Links 20. Rossaint R, Hahn S-M, Pappert D, et al. Influence of mixed venous PO 2 and inspired oxygen fraction on intrapulmonary shunt in patients with severe ARDS. J Appl Physiol 1995;78:1531–1536. Bibliographic Links 21. Lele A, Mirski MA, Stevens RD. Spurious hypoxemia. Crit Care Med 2005;33: 1854–1856. Ovid Full TextBibliographic Links 22. Fox MJ, Brody JS, Weintraub LR. Leukocyte larceny: a cause of spurious hypoxemia. Am J Med 1979;67:742–746. Bibliographic Links

Hypercapnia 23. Weinberger SE, Schwartzstein RM, Weiss JW. Hypercapnia. N Engl J Med 1989;321:1223–1230. Bibliographic Links 24. Talpers SS, Romberger DJ, Bunce SB, et al. Nutritionally associated increased carbon dioxide production. Chest 1992;102:551–555. Bibliographic Links 25. Gray BA, Blalock JM. Interpretation of the alveolar-arterial oxygen difference in patients with hypercapnia. Am Rev Respir Dis 1991;143:4–8. Bibliographic Links

Chapter 20 Oximetry and Capnography The noninvasive detection of blood gases (PO 2, PCO2) using optical and colorimetric techniques (1,2) is the most significant and useful advance in critical care monitoring in the last quarter century. This chapter describes the monitoring techniques that have become an integral part of daily patient care in the ICU (and in most other areas of the hospital as well). Despite the popularity of these techniques, surveys reveal that 95% of ICU staff members have little or no understanding of how the techniques work (3).

Oximetry All atoms and molecules absorb specific wavelengths of light (this is the source of color in the lighted world). This property is the basis for an optical technique known as spectrophotometry, which transmits light of specific wavelengths through a medium to determine the molecular composition of the medium. The absorption of light as it passes through a medium is proportional to the concentration of the substance that absorbs the light and the length of the path that the light travels (this is known as the Lambert-Beer Law). The application of this principle to the detection of hemoglobin in its different forms is known as oximetry.

Optical Recognition of Hemoglobin Hemoglobin (like all proteins) changes its structural configuration when it participates in a chemical reaction, and each of the configurations has a distinct pattern of light absorption. The patterns of light absorption for the different forms of hemoglobin are shown in Figure 20.1. Four different forms of hemoglobin are represented in the figure: oxygenated hemoglobin (HbO2) deoxygenated hemoglobin (Hb), methemoglobin (met Hb) and carboxyhemoglobin (COHb). Comparing the oxygenated and deoxygenated forms of hemoglobin (HbO 2 and Hb) shows that, in the red region of the light spectrum (660 nm), HbO2 does not absorb light P.386 as well as Hb (this is why oxygenated blood is more intensely red than deoxygenated blood), while in the infrared region (940 nm), the opposite is true, and HbO 2 absorbs light more effectively than Hb.

View Figure

Figure 20.1 The absorption spectrum for the different forms of hemoglobin: oxygenated hemoglobin (HbO 2), deoxygenated hemoglobin (Hb), carboxyhemoglobin (COHb), and methemoglobin (metHb) The vertical lines represent the two wavelengths of light (660 nm and 940 nm) used by pulse oximeters. (Adapted from Barker SJ, Tremper KK. Pulse oximetry: applications and limitations. Internat Anesthesiol Clin 1987;25:155)

Because methemoglobin (metHb) and carboxyhemoglobin (COHb) make up less than 5% of the total hemoglobin pool in most situations (2,4), the transmission of light at 660 nm through a blood sample is determined by the amount of HbO2 in the sample, while light transmission at 940 nm is determined by amount of Hb in the sample. The amount of HbO 2 can then be compared to the total amount of hemoglobin (HbO 2 + Hb) to express the fraction of the hemoglobin pool that is in the oxygenated form. This is known as % saturation, and is derived in Equation 20.1 below. This is how most bedside oximeters operate, using two wavelengths of light (660 and 940 nm), and expressing the oxygenated hemoglobin as a percentage of the total hemoglobin.

Early Oximeters The first bedside oximeters used probes that were clamped onto an earlobe. A light emitting device on one side of the probe sent red and P.387 infrared light through the earlobe to a photodetector on the other side, which amplified the transmitted light. The intention was to measure the oxygenated hemoglobin in the small arterioles within the earlobe. These devices suffered from two shortcomings: (1) the transmission of light was affected by factors other than hemoglobin (e.g., skin pigments), and (2) it was not possible to distinguish between oxyhemoglobin saturation in arteries and veins.

Pulse Oximetry The introduction of pulse oximetry in the mid-1970s eliminated many of the problems that plagued the early oximeters. The basic operation of a pulse oximeter is shown in Figure 20.2. The probes on pulse oximeters are shaped like sleeves that are placed around a

finger. One side of the probe has a phototransmitter that emits monochromatic light at wavelengths of 660 and 940 nm. The light travels through the tissues of the finger to reach a photodetector on the other side. The unique feature of pulse oximeters is the photodetector, which amplifies only light of alternating intensity. (This is analogous to an AC amplifier, which amplifies only alternating-current impulses.) Light that strikes a pulsating artery will develop phasic changes in intensity and will be amplified by the photodetector, while light that passes through non-pulsatile tissue will be blocked by the photodetector. This allows pulse oximeters to detect only the hemoglobin in pulsating arteries, and it reduces or eliminates errors created by light absorption in non-pulsatile structures like connective tissue and veins. Figure 20.2 The principle of pulse oximetry. The photodetector senses only light of alternating intensity (analogous to an AC amplifier).

View Figure

TABLE 20.1 Variability in Oximetry and Capnometry Recordings

Study Parameters

SpO 2*



Time period

60 min

120 min

60 min

Mean variation



2 mm Hg

Range of variation



0–7 mm Hg

Clinically stable patients. 95% of measurements obtained during mechanical ventilation. *From Reference 6. **From Reference 19.


Accuracy At clinically acceptable levels of arterial oxygenation (SaO 2 above 70%), the O2 saturation recorded by pulse oximeters (SpO 2) differs by less than 3% from the actual SaO2 (4,5). SpO2 also shows a high degree of precision (consistency of repeated measurements). This is demonstrated in Table 20.1 (6), which shows that SpO2 varies by = 2% in most patients who are clinically stable.

Carbon Monoxide Intoxication Carbon monoxide displaces oxygen from the iron binding sites in hemoglobin, so carbon monoxide intoxication will increase carboxyhemoglobin (COHb) levels and decrease oxyhemoglobin (HbO2) levels. This should be evident as a decreased arterial O 2 saturation (SaO2). However, as demonstrated in Figure 20.1, light absorption at 660 nm is similar for carboxyhemoglobin (COHb) and oxygenated hemoglobin (HbO 2). This means that pulse oximeters will mistake COHb for HbO 2, and the SpO2 will be higher than the actual SaO2. The difference between the SpO 2 and the SaO2 (SpO2 - SaO2), the pulse oximetry gap, is equivalent to the COHb level (7). Because the SpO 2 overestimates the SaO 2 when COHb levels are elevated, pulse oximetry is unreliable for detecting hypoxemia in carbon monoxide intoxication. When carbon monoxide intoxication is suspected, an arterial blood sample should be sent to the clinical laboratory for a direct measurement of the COHb level. (Clinical laboratories have multiple-wavelength spectrophotometers that can more precisely measure the different forms of hemoglobin in the blood.)

Methemoglobinemia The oxidized iron moieties in methemoglobin do not carry oxygen effectively, so accumulation of methemoglobin (metHb) will decrease the SaO 2. Pulse oximetry overestimates the SaO2 (SpO2 > SaO2), and the SpO2 rarely falls below 85% in methemoglobinemia despite much lower levels of SaO 2 (8). Thus, pulse oximetry should not be used when methemoglobinemia is suspected. Accurate measurements of metHb, and SaO 2 require the more sophisticated spectrophotometers in clinical laboratories. P.389

Hypotension Although pulse oximetry is based on the presence of pulsatile blood flow, SpO 2 is an accurate reflection of SaO 2 down to blood pressures as low as 30 mm Hg (9). Damped pulsations also do not affect the accuracy of fingertip SpO 2 recordings taken distal to a cannulated radial artery (10). For situations where fingertip SpO2 recordings may be problematic because of severely reduced peripheral blood flow, specialized oximeter sensors are available that can be placed on the forehead. These sensors differ from the fingertip sensors because they record light that is reflected back to the skin surface (reflectance spectrophotometry). Forehead sensors respond much more rapidly to changes in SpO 2 than fingertip sensors (11), and they should gain popularity as a suitable alternative to the traditional fingertip


Anemia In the absence of hypoxemia, pulse oximetry is accurate down to hemoglobin levels as low as 2 to 3 g/dL (12). With lesser degrees of anemia (Hb between 2.5 and 9 g/dL), SpO2 underestimates SaO2 by only 0.5% (12).

Pigments The effects of dark skin pigmentation on the SpO2 has varied in different reports. In one study, the SpO2 was spuriously low in patients with dark skin (13), while in another study, the SpO 2 was spuriously high (SpO2 - SaO2 = 3.5%) when the SaO 2 was less than 70% (14). Fingernail polish has a small effect on the SpO 2 when the color is black or brown (SpO2 2% less than SaO2), but this effect can be eliminated by placing the probes on the side of the finger (15). The largest pigment effect is produced by methylene blue, which can produce a 65% decrease in SpO 2 when injected intravenously (4). Because methylene blue is used to treat methemoglobinemia, this is another reason to avoid pulse oximetry in patients with methemoglobinemia.

Detecting Hypoventilation Clinical studies have shown that the SpO2 can be a sensitive marker of inadequate ventilation (a low PaO2) when patients are breathing room air, but not when they are breathing supplemental oxygen (16). This is explained by the shape of the oxyhemoglobin dissociation curve. When SpO2 (or SaO2) exceeds 90% (PaO2 > 60 mm Hg), the curve begins to flatten, and larger changes in PaO 2 are accompanied by smaller changes in SpO2. Breathing supplemental oxygen will push the SpO 2 further out onto the flat part of the oxyhemoglobin dissociation curve (the SpO 2 is often >98% during supplemental O2 breathing), where relatively large changes in PaO2 are accompanied by minor changes in the SpO 2. There is a tendency to use supplemental oxygen routinely in the ICU (and the postanesthesia recovery unit) even when the SpO 2 exceeds 90%. Because there is no documented benefit to increasing the SaO2 far above 90%, supplemental O2 can be safely withheld if the SpO2 is P.390 92% or higher on room air. This practice will limit unnecessary oxygen administration (to limit the toxic effects of oxygen), and will preserve the sensitivity of pulse oximetry in detecting inadequate ventilation.

When to Use Pulse Oximetry Considering that the SpO2 has been called the fifth vital sign, it might be more appropriate to consider when not to use pulse oximetry. In short, pulse oximetry is indicated in any situation where monitoring arterial oxygenation is considered important. In critically ill patients, at least 15 clinical studies have shown that continuous monitoring of SpO 2 with pulse oximetry is superior to periodic blood gas measurements for detecting episodes of

significant hypoxemia (4). The combination of pulse oximetry and end-tidal CO 2 monitoring (described next) should largely replace the more expensive and painful method of arterial blood gas measurements.

Venous Oximetry The O2 saturation in the superior vena cava or pulmonary artery can be monitored continuously with specialized catheters that emit red and infrared light from the catheter tip and record the light reflected back from the hemoglobin in circulating erythrocytes (see Figure 20.3) (17). This technique of reflectance spectrophotometry is a variant of the transmission spectrophotometry used by fingertip probes for pulse oximetry. Most venous oximetry systems process and display the venous O 2 saturation every 5 seconds. Figure 20.3 Continuous measurement of O2 saturation of hemoglobin in mixed venous blood (SvO2) using reflectance spectrophotometry.

View Figure


Mixed Venous O2 Saturation The interpretation of mixed venous O 2 saturation (SvO2) is described in Chapter 11. The SvO2 is measured in pulmonary artery blood, and is a marker of the balance between whole-body O2 delivery (DO2) and O2 consumption (VO2): SvO2 = DO2/VO2. A decrease in SvO2 below the normal range of 70 to 75% identifies a state of inadequate O 2 delivery relative to O2 consumption that could be the result of a decreased DO 2 (from low cardiac output, anemia, or hypoxemia) or an increased VO2 (from hypermetabolism). (The determinants of DO 2 and VO2 are described in Chapter 2.) The continuous measurement of SvO 2 using specialized pulmonary artery catheters is accurate to within 1.5% of the SvO 2 measured in the clinical laboratory (18). Despite this acceptable accuracy, SvO2 can vary considerably without an apparent change in hemodynamic status. The spontaneous variability of SvO 2 is shown in Table 20.1. The average variation over a 2-hour period is 6%, but can be as high as 19% (19). As a general rule, a greater than 5% variation in SvO2 that persists for longer than 10 minutes is considered a significant change (20).

Central Venous O2 Saturation

Continuous monitoring of the central venous O 2 saturation (ScvO2) is achieved with specialized central venous catheters placed in the superior vena cava. The ScvO 2 tends to be slightly lower than the SvO 2, and this difference is magnified in the presence of circulatory shock (17). Single measurements of ScvO2 can differ from SvO2 by as much as 10%, but the difference is reduced (to within 5%) when multiple measurements are obtained (21). The ScvO2 seems most valuable in identifying trends in the balance between DO 2 and VO 2. Central venous oximetry is gaining popularity over mixed venous oximetry because of the cost and morbidity associated with pulmonary artery catheters. Recent guidelines for the early management of patients with severe sepsis and septic shock includes an ScvO 2 of greater than 70% as a therapeutic end-point (22).

Dual Oximetry The predictive value of SvO2 or ScvO2 can be increased by adding the SaO 2 measured by pulse oximetry (SpO2). This provides a continuous measure of (SaO2 - SvO2), which is equivalent to the O2 extraction from capillary blood (23). The determinants of the (SaO 2 SvO2) can be derived using the determinants of DO 2 and VO2 (which are described in Chapter 2): An increase in (SaO 2 - SvO2) above the normal range of 20 to 30% can be the result of increased VO 2 (hypermetabolism), a reduced Q (low cardiac output) or a reduced Hb (anemia). The (SaO 2 - SvO2) may be P.392 more valuable as a marker of tissue dysoxia (defined as a state of oxygen-limited metabolism) or impending dysoxia. For example, in the presence of anemia or a low cardiac output, an increase in (SaO2 - SvO2) to its maximum level of 50 to 60% indicates that the tissues are no longer able to compensate for further reductions on Hb or cardiac output, which means there is a risk for tissue anaerobiasis. Thus, in a patient with progressive anemia, an (SaO2 - SvO2) of 50 to 60% could be used as an indication for transfusion (transfusion trigger).

Capnometry Capnometry is the measurement of CO2 in exhaled gas. This can be achieved with a colorimetric technique, or with infrared spectrophotometry. Both methods are described below.

Colorimetric CO2 Detection The colorimetric detection of CO 2 in exhaled gas is a quick and simple method of determining if an endotracheal tube has been placed in the lungs ( 24,25). This is recommended as a standard practice following attempted intubation because auscultation for breath sounds is an unreliable method of determining if an

endotracheal tube is in the esophagus or lungs (26). The most popular colorimetric CO 2 detector in clinical practice is illustrated in Figure 20.4. This device has two ports for attachment: one to the endotracheal tube and the other end to an inflatable resuscitation bag. The central area of the device contains filter paper that is impregnated with a pH-sensitive indicator that changes color as a function of pH. When exhaled gas passes over the filter paper, the CO 2 in the gas is hydrated by a liquid film on the filter paper, and the resulting pH is detected by a color change. The outer perimeter of the chemical reaction area contains color-coded sections indicating the concentrations of exhaled CO2 associated with each color change. A purple color indicates 1.5, a creatine kinase (CPK) >5,000 IU/L, a base deficit =4, and myoglobin in the urine (23). The serum creatinine is not an accurate index of renal function in rhabdomyolysis because the enhanced creatine release from skeletal muscle adds to the serum creatinine. The principal conditions that predispose to renal injury are hypovolemia and acidosis.

Myoglobin in Urine Myoglobin can be detected in urine with the orthotoluidine dipstick reaction (Hemastix) that is used to detect occult blood in urine. If the test is positive, the urine should be centrifuged (to separate erythrocytes) and the supernatant should be passed through a

micropore filter (to remove hemoglobin). A persistently positive test after these measures is evidence of myoglobin in urine. The presence of myoglobin in urine does not identify patients with a high risk of renal injury, but the absence of myoglobin in urine identifies patients with a low risk of renal injury (22).

Management The plasma levels of potassium and phosphate must be monitored carefully in rhabdomyolysis because these electrolytes are released by injured skeletal muscle and the concentration in plasma can increase dramatically, especially when renal function is impaired. Aggressive volume resuscitation to prevent hypovolemia and maintain renal blood flow is one of the most effective measures for preventing or limiting the renal injury in rhabdomyolysis. Alkalinizing the urine can also help to limit the renal injury, but this is difficult to accomplish and is often not necessary. About 30% of patients who develop myoglobinuric renal failure will require dialysis (22).

Renal Replacement Therapy About 70% of patients with acute renal failure will require some form of renal replacement therapy (RRT). The usually indications for RRT are volume overload, uremic encephalopathy, and difficult-to-control hyperkalemia and metabolic acidosis. There is a growing body of RRT techniques, including hemodialysis, hemofiltration, hemodiafiltration, high flux dialysis, and plasmafiltration. Each employs a different method of water and solute transport. The presentation that follows is limited to the techniques of hemodialysis and hemofiltration. Figure 31.3 A double lumen venous catheter used for hemodialysis. The internal diameter of each lumen is about twice the diameter of each lumen in a triple-lumen central venous catheter (see Figure 6.4).

View Figure


Hemodialysis Hemodialysis removes solutes by diffusion, which is driven by the concentration gradient of the solutes across a semipermeable membrane. To maintain this concentration

gradient, blood and dialysis fluid are driven in opposite directions across the diffusional barrier (dialysis membrane): this technique is known as countercurrent exchange (24). A blood pump is used to move blood in one direction across the dialysis membrane at a rate of 200 to 300 mL/min. The dialysis fluid on the other side of the membrane moves twice as fast, or at 500 to 800 mL/min (24).

Vascular Access A large-bore, double-lumen vascular catheter like the one shown in Figure 31.3 is required to perform intermittent hemodialysis. Each lumen of the catheter in this figure has a diameter that is roughly twice the diameter of each lumen in a standard central venous catheter (see the triple-lumen catheter in Figure 6.4 and compare lumen sizes using Table 6.1). Because flow through rigid tubes varies directly with the fourth power of the radius (see the Hagen-Poiseuille equation in Figure 1.6), a catheter lumen that is doubled in size will allow (2 4 = 16) a 16-fold greater flow P.590 rate. The large-bore catheter in Figure 31.3 is thus well-suited for the high flow needed to perform intermittent hemodialysis. Blood is withdrawn from one lumen of the catheter, pumped through the dialysis chamber, and then returned to the patient through the other lumen. Figure 31.4 The technique of continuous arteriovenous hemofiltration (CAVH).

View Figure

The large-bore dialysis catheters are placed in either the internal jugular vein or the femoral vein. The subclavian vein is not recommended because there is a high incidence of vascular stenosis and this makes the ipsilateral arm veins unsuitable for chronic dialysis access if renal function does not recover (25). The internal jugular vein is preferred because of the risk for venous thrombosis with femoral vein cannulation (see Table 6.3), but awake patients are intolerant of the limited neck mobility associated with cannulation of the internal jugular vein with these large-bore catheters. P.591

Benefits and Risks The benefit of hemodialysis is rapid clearance of solutes. Only a few hours of dialysis is needed to remove a day's worth of accumulated nitrogenous waste. The disadvantage of dialysis is the need to maintain a blood flow of at least 300 mL/min through the dialysis chamber. This creates a risk of hypotension, which occurs in about one-third of intermittent hemodialysis treatments (24).

Hemofiltration Whereas hemodialysis removes solutes by diffusion, hemofiltration uses convection for solute transport. Convection is a method where a solute-containing fluid is driven across a permeable membrane by exerting a pressure difference across the membrane. The solutes are cleared by the movement of the fluid across the membrane. Since the fluid “drags” the solute across the membrane, this method of solute transport is known as solvent drag (24). Because the removal of solutes by convection is relatively slow, hemofiltration is performed continuously.

Continuous Hemofiltration One technique of continuous hemofiltration is illustrated in Figure 31.4. In this case, the hemofilter is placed between an artery and a vein, and the technique is called continuous arteriovenous hemofiltration (CAVH). No pumps are required for CAVH. The arteriovenous pressure difference is the pressure gradient for flow through the filter. The pressure gradient for water movement across the filter is the mean blood pressure on one side, and the vertical distance between the filter and the ultrafiltrate collection bag on the other side. If the collection bag is lowered, the filtration pressure will increase. Hemofiltration can remove large volumes of fluid (up to 3 liters per hour) so a replacement fluid is needed to prevent hypovolemia. This is shown in the CAVH circuit in Figure 31.4. The replacement fluid also decreases the plasma concentration of waste products that are removed in the ultrafiltrate. That is, the concentration of solutes in the ultrafiltrate is the same as in the blood, so the plasma concentration of waste products will not decrease unless a waste-free fluid is used to replace the fluid that is removed.

Benefits Since hemofiltration is driven by pressure and not flow, high flow rates are not needed, and there is a much less risk of hypotension with CAVH. The removal of solutes is also more gradual and more physiological with continuous hemofiltration. One shortcoming of CAVH is that it is not suitable for use in patients with hypotension. The technique of continuous venovenous hemofiltration (CVVH) uses a pump to generate a pressure, and can be used in hypotensive patients. In general, the benefits of continuous renal replacement therapy are leading to a gradual disappearance of intermittent hemodialysis in the ICU. P.592

A Final Word A sudden and precipitous drop in urine output is rarely a sign of simple dehydration that

will be corrected with fluids. Instead, it is usually an ominous sign of failure involving of one of the major organ systems in the body. In the setting of sepsis, the appearance of oliguria usually heralds the beginning of multiple organ failure, which often has a fatal outcome. This condition of multiorgan failure, which can be viewed as the gradual process of dying (the more organs that fail, the closer the patient is to death), is described in Chapter 40.

References Introduction 1. Uchino S, Kellum JA, Bellomo R, et al. Acute renal failure in critically ill patients: a multinational, multicenter study. JAMA 2005;294:813–818. Ovid Full TextBibliographic Links

Reviews 2. Bellomo R. Defining, quantifying, and classifying acute renal failure. Crit Care Clin 2005;21:223–237. Bibliographic Links 3. Abernathy VE, Lieberthal W. Acute renal failure in the critically ill patient. Crit Care Clin 2002;18:203–222. Bibliographic Links 4. Singri N, Ahya SN, Levin ML. Acute renal failure. JAMA 2003;289:747–751. Ovid Full TextBibliographic Links 5. Klahr S, Miller SB. Acute oliguria. N Engl J Med 1998;338:671–675. Ovid Full TextBibliographic Links

Evaluation of Oliguria 6. Dellinger RP, Carlet JM, Masur H, et al. Surviving sepsis campaign guidelines for management of severe sepsis and septic shock. Crit Care Med 2004; 32:858–873. Ovid Full TextBibliographic Links 7. Steiner RW. Interpreting the fractional excretion of sodium. Am J Med 1984; 77:699–702. Bibliographic Links 8. Cockroft DW, Gault MN. Prediction of creatinine clearance from serum creatinine.

Nephron 1976;16:31–41. Bibliographic Links

Initial Management 9. Vincent J-L, Gerlach H. Fluid resuscitation in severe sepsis and septic shock: an evidence-based review. Crit Care Med 2004;32(suppl):S451–S454. Ovid Full TextBibliographic Links 10. Kellum JA, Decker JM. Use of dopamine in acute renal failure: a meta-analysis. Crit Care Med 2001;29:1526–1531. Ovid Full TextBibliographic Links 11. Holmes CL, Walley KR. Bad medicine: low-dose dopamine in the ICU. Chest 2003;123:1266–1275. Full TextBibliographic Links 12. Brater DC, Anderson SA, Brown-Cartwright D. Response to furosemide in chronic renal insufficiency: rationale for limited doses. Clin Pharmacol Ther 1986;40:134–139. Bibliographic Links P.593

Specific Renal Disorders 13. Schrier RW, Wang W. Acute renal failure and sepsis. N Engl J Med 2004;351: 159–169. Ovid Full TextBibliographic Links 14. Balk RA. Pathogenesis and management of multiple organ dysfunction or failure in severe sepsis and septic shock. Crit Care Clin 2000;16:337–352. Bibliographic Links 15. Pinsky MR, Vincent J-L, Deviere J, et al. Serum cytokine levels in human septic shock: relation to multiple-system organ failure and mortality. Chest 1993;103:565–575. Bibliographic Links 16. McCullough PA, Soman S. Contrast-induced nephropathy. Crit Care Clin 2005;21:261–280. Bibliographic Links 17. Liu R, Nair D, Ix J, et al. N-acetylcysteine for the prevention of contrast-induced

nephropathy. A systematic review and meta-analysis. J Gen Intern Med 2005;20:193–200. Ovid Full TextBibliographic Links 18. Marenzi G, Assanelli E, Marana I, et al. N-acetylcysteine and contrast-induced nephropathy in primary angioplasty. N Engl J Med 2006;354:2772–2782.

19. Taber SS, Mueller BA. Drug-associated renal dysfunction. Crit Care Clin 2006;22:357–374. Bibliographic Links 20. Ten RM, Torres VE, Millner DS, et al. Acute interstitial nephritis. Mayo Clin Proc 1988;3:921–930. Bibliographic Links 21. Beetham R. Biochemical investigation of suspected rhabdomyolysis. Ann Clin Biochem 2000; 2000:37:581–587.

22. Sharp LS, Rozycki GS, Feliciano DV. Rhabdomyolysis and secondary renal failure in critically ill surgical patients. Am J Surg 2004;188:801–806. Bibliographic Links 23. Visweswaran P, Guntupalli J. Rhabdomyolysis. Crit Care Clin 1999;15:415–428. Bibliographic Links

Renal Replacement Therapy 24. O'Reilly P, Tolwani A. Renal replacement therapy III. IHD, CRRT, SLED. Crit Care Clin 2005;21:367–378. Bibliographic Links 25. Hernandez D, Diaz F, Rufino M, et al. Subclavian vascular access stenosis in dialysis patients: natural history and risk factors. J Am Soc Nephrol 1998;9: 1507–1511. Bibliographic Links

Chapter 32 Hypertonic and Hypotonic Conditions This chapter describes the diagnosis and management of conditions associated with abnormalities in total body water. These conditions typically present with abnormalities in the plasma sodium concentration (hypernatremia and hyponatremia) that are sometimes mistaken as problems in sodium balance (1,2). This chapter presents a very simple approach to hypernatremia and hyponatremia based on a clinical assessment of the extracellular volume. The very first part of the chapter contains a quick review of the determinants of water movement between fluid compartments.

Basic Concepts The following is a description of the forces that determine the movement of water between the intracellular and extracellular fluid compartments.

Osmotic Activity The activity (concentration) of solute particles in a solution is inversely related to the activity (concentration) of water molecules in the solution. The solute activity in a solution is also called the osmotic activity and is expressed in osmoles (osm). The total osmotic activity in a solution is the sum of the individual osmotic activities of all the solute particles in the solution. For monovalent ions, the osmotic activity in milliosmoles (mOsm) per unit volume is equivalent to the concentration of the ions in milliequivalents (mEq) per unit volume. Thus, the osmotic activity in isotonic saline (0.9% sodium chloride) is as follows:

P.596 Osmolarity is the osmotic activity per volume of solution (solutes plus water) and is expressed as mOsm/L (3,4). Osmolality is the osmotic activity per volume of water and is expressed as mOsm/kg H2O. The osmotic activity of body fluids usually is expressed in relation to the volume of water (i.e., osmolality). However, the volume of water in body fluids is far greater than the volume of solutes, so there is little difference between the osmolality and osmolarity of body fluids. Thus, the terms osmolality and osmolarity can be used interchangeably to describe the osmotic activity in body fluids.


When two solutions are separated by a membrane that allows the passage of water but not solutes, the water passes from the solution with the lower osmotic activity to the solution with the higher osmotic activity. The relative osmotic activity in the two solutions is called the effective osmolality, or tonicity. The solution with the higher osmolality is described as hypertonic, and the solution with the lower osmolality is described as hypotonic. Thus, the tendency for water to move into and out of cells is determined by the relative osmolality (tonicity) of the intracellular and extracellular fluids. When the membrane separating two fluids is permeable to both solutes and water, and a solute is added to one of the fluids, the solute equilibrates fully across the membrane. In this situation, the solute increases the osmolality of both fluids, but there will be no movement of water between compartments (because there is no difference in osmolality between the two compartments). A solute that behaves in this manner is urea, which is freely permeable across cell membranes. Therefore, an increase in the urea concentration in extracellular fluid (i.e., an increase in the blood urea nitrogen or BUN) will increase the osmolality of the extracellular fluid, but this does not draw water out of cells because urea does not create a difference in osmolality between extracellular and intracellular fluid. Thus, azotemia (increased BUN) is a hyperosmotic condition, but not a hypertonic condition.

Plasma Osmolality The osmolality of the extracellular fluids can be measured in the clinical laboratory using the freezing point of plasma (a solution containing 1 osm/L will freeze at 21.86°C). This is the freezing point depression method for measuring osmolality. The osmolality of the extracellular fluids can also be calculated using the concentrations of sodium, chloride, glucose, and urea in plasma (these are the major solutes in extracellular fluid). The calculation below uses a plasma sodium (Na 1) of 140 mEq/L, a plasma glucose of 90 mg/dL, and a BUN of 14 mg/dL (3,5).

P.597 The sodium concentration is doubled to include the osmotic contribution of chloride. The serum glucose and urea are measured in milligrams per deciliter, and the factors 18 and 2.8 (the atomic weights divided by 10) are used to convert mg/dL to mOsm/kg H 2O.

Osmolal Gap Because solutes other than sodium, chloride, glucose, and urea are present in the extracellular fluid, the measured plasma osmolality will be greater than the calculated plasma osmolality. This osmolar gap (i.e., the difference between the measured and calculated plasma osmolality) is normally as much as 10 mOsm/kg H 2O (3,5). An increase in the osmolar gap occurs when certain toxins (e.g., ethanol, methanol, ethylene glycol, or the unidentified toxins that accumulate in renal failure) are in the extracellular fluid (6). Therefore, the osmolar gap has been proposed as a screening test for

identifying the presence of toxins in the extracellular fluid. In the case of renal failure, the osmolar gap has been recommended as a reliable test for distinguishing acute from chronic renal failure: the osmolar gap is expected to be normal in acute renal failure and elevated in chronic renal failure (7). In reality, the osmolar gap is used infrequently.

Plasma Tonicity Because urea passes freely across cell membranes, the effective osmolality or tonicity of the extracellular fluid can be calculated by eliminating urea (BUN) from the plasma osmolality equation.

Because the concentration of urea contributes little to the total solute concentration in extracellular fluids, there is little difference between the osmolality and tonicity of the extracellular fluid. This equation establishes the plasma sodium concentration as the principal determinant of the effective osmolality of extracellular fluid. Because the effective osmolality determines the tendency for water to move into and out of cells, the plasma sodium concentration is the principal determinant of the relative volumes of the intracellular and extracellular fluids.

Hypernatremia The normal plasma (serum) sodium concentration is 135 to 145 mEq/L. Therefore, hypernatremia (i.e., a serum sodium concentration above 145 mEq/L) can be the result of loss of fluid that has a sodium P.598 concentration below 135 mEq/L (hypotonic fluid loss) or gain of fluid that has a sodium concentration above 145 mEq/L (hypertonic fluid gain). Each of these conditions can be identified by assessing the state of the extracellular volume as shown in Table 32.1 (1,8,9). TABLE 32.1 Change in Total Body Sodium and Water in Hypernatremia and Hyponatremia




Extracellular Volume

Total Body Sodium

Free Water

















Extracellular Volume If invasive hemodynamic monitoring is available, the state of the intravascular volume can be evaluated by the relationship between the cardiac filling pressures and the cardiac output. For example, the combination of reduced cardiac filling pressures and a low cardiac output is evidence of hypovolemia (see the section on Hemodynamic Subsets in Chapter 9). In the absence of hypoproteinemia (which shifts fluids from the intravascular to extravascular space), the state of the intravascular volume can be used as a reflection of the state of the extracellular volume (ECV). If invasive hemodynamic monitoring is not available, the evaluation of hypovolemia described in Chapter 12 can be used to detect a decreased ECV. The clinical detection of an increased ECV can be difficult because the absence of edema does not exclude the presence of a high ECV (edema may not be apparent until the ECV has increased 4 to 5 liters) and the presence of edema can be misleading because edema in ICU patients can be the result of immobility, hypoalbuminemia, or venous congestion from high intrathoracic pressures (in ventilator-dependent patients). Once the state of the ECV is determined, the strategies shown in Figure 32.1 can be applied. Low ECV indicates loss of hypotonic fluids. Common causes are excessive diuresis, vomiting, and diarrhea. The management strategy is to replace the sodium deficit quickly (to maintain plasma volume) and to replace the free water deficit slowly (to prevent intracellular overhydration). Normal ECV indicates a net loss of free water. This can be seen in diabetes insipidus, or

when loss of hypotonic fluids (e.g., diuresis) is P.599 treated by replacement with isotonic saline in a 1:1 volume-to-volume ratio. The management strategy is to replace the free water deficit slowly (to prevent intracellular overhydration). Figure 32.1 Management of hypernatremia based on the extracellular volume.

View Figure

High ECV indicates a gain of hypertonic fluids. This is seen with aggressive use of hypertonic saline or sodium bicarbonate solutions. The management strategy is to induce sodium loss in the urine with diuresis and to replace the urine volume loss with fluids that are hypotonic to the urine. Each of these conditions is described in more detail in the following sections.

Hypovolemic Hypernatremia The most common cause of hypernatremia is loss of hypotonic body fluids. The concentration of sodium in body fluids that are commonly lost is shown in Table 32.2. With the exception of small bowel and pancreatic secretions, loss of any of these body fluids will result in hypernatremia.

Consequences All of the body fluids listed in Table 32.2 contain sodium, so the loss of these fluids will be accompanied by deficits in total body sodium as well P.600 as total body water (TBW). The sodium deficits predispose to hypovolemia, whereas the free water deficits predispose to hypertonicity in the extracellular fluids. Therefore, the two consequences of hypotonic fluid loss are hypovolemia and hypertonicity.

TABLE 32.2 Sodium Concentration in Body Fluids

Fluids Commonly Lost Urine*

Sodium Concentration (mEq/L) 20 mEq/L

Renal failure

20 mEq/L

The urine sodium can be misleading if the patient is also receiving diuretics (which are commonly used in these conditions). The clinical picture is usually helpful, although these conditions can co-exist in critically ill patients.

Hyponatremic Encephalopathy The most feared complication of hyponatremia is a life-threatening metabolic encephalopathy that is often associated with cerebral edema, increased intracranial pressure, and seizures, and can be accompanied by the adult respiratory distress syndrome (27,28,29). Severe cases can progress to respiratory arrest. Correction of the hyponatremia can also be associated with an encephalopathy that is characterized by demyelinating lesions, pituitary damage, and oculomotor nerve palsies (28). This is usually seen when the sodium concentration is corrected too rapidly. A specific demyelinating disorder known as central pontine myelinolysis has also been attributed to rapid correction of hyponatremia (30). These conditions can be irreversible and even fatal, and the next section contains some recommendations to limit the risk of central nervous system injury.

Management Strategies The management of hyponatremia is determined by the state of the ECV (i.e., low, normal, or high) and by the presence or absence of neurological symptoms. Symptomatic hyponatremia requires more aggressive corrective therapy than asymptomatic hyponatremia. However, to limit the risk of a demyelinating encephalopathy, the rate of rise in plasma sodium should not exceed 0.5 mEq/L per hour and the final plasma sodium P.608 concentration should not exceed 130 mEq/L (27). The general management strategies based on the ECV are as follows: Low ECV: Infuse hypertonic saline (3% NaCl) in symptomatic patients, and isotonic saline in asymptomatic patients. Normal ECV: Combine furosemide diuresis with infusion of hypertonic saline in symptomatic patients, or isotonic saline in asymptomatic patients. High ECV: Use furosemide-induced diuresis in asymptomatic patients. In symptomatic

patients, combine furosemide diuresis with judicious use of hypertonic saline.

Sodium Replacement When corrective therapy requires the infusion of isotonic saline or hypertonic saline, the replacement therapy can be guided by the calculated sodium deficit. This is determined as follows (using a plasma sodium of 130 mEq/L as the desired end-point of replacement therapy): The normal TBW (in liters) is 60% of the lean body weight (in kg) in men, and 50% of the lean body weight in women. Thus, for a 60 kg woman with a plasma sodium of 120 mEq/L, the sodium deficit will be 0.5 × 60 × (130 - 120) = 300 mEq. Because 3% sodium chloride contains 513 mEq of sodium per liter, the volume of hypertonic saline needed to correct a sodium deficit of 300 mEq will be 300/513 = 585 mL. Using a maximum rate of rise of 0.5 mEq/L per hour for the plasma sodium (to limit the risk of a demyelinating encephalopathy), the sodium concentration deficit of 10 mEq/L in the previous example should be corrected over at least 20 hours. Thus, the maximum rate of hypertonic fluid administration will be 585/20 = 29 mL/hour. If isotonic saline is used for sodium replacement, the replacement volume will be 3.3 times the replacement volume of the hypertonic 3% saline solution.

A Final Word To design an effective approach to hypernatremia and hyponatremia, it is essential to understand that these conditions are the result of a problem with water balance more than sodium balance. The approach in this chapter shows you how to identify the problem with water and sodium balance in any patient using one determination: i.e., an assessment of the extracellular volume.

References Reviews 1. Verbalis JG. Disorders of body water homeostasis. Best Pract Res Clin Endocrinol Metab 2003;17:471–503. Bibliographic Links P.609 2. Rose BD, Post TW. The total body water and the plasma sodium concentration. In: Clinical physiology of acid-base and electrolyte disorders. 5th ed. New York: McGraw-Hill, 2001: 241–257.

Basic Concepts

3. Gennari FJ. Current concepts: serum osmolality: uses and limitations. N Engl J Med 1984;310:102–105. Bibliographic Links 4. Erstad BL. Osmolality and osmolarity: narrowing the terminology gap. Pharmacotherapy 2003;23:1085–1086. Bibliographic Links 5. Turchin A, Seifter JL, Seely EW. Clinical problem-solving: mind the gap. N Engl J Med 2003;349:1465–1469. Ovid Full TextBibliographic Links 6. Purssell RA, Lynd LD, Koga Y. The use of the osmole gap as a screening test for the presence of exogenous substances. Toxicol Rev 2004;23:189–202.

7. Sklar AH, Linas SL. The osmolal gap in renal failure. Ann Intern Med 1983;98:481–482. Full TextBibliographic Links

Hypernatremia 8. Adrogue HJ, Madias NE. Hypernatremia. N Engl J Med 2000;342:1493–1499. Ovid Full TextBibliographic Links 9. McGee S, Abernethy WB 3rd, Simel DL. The rational clinical examination: Is this patient hypovolemic? JAMA 1999;281:1022–1029.

10. Arieff AI, Ayus JC. Strategies for diagnosing and managing hypernatremic encephalopathy. J Crit Illness 1996;11:720–727.

11. Rose BD, Post TW. Hyperosmolal states: hyperglycemia. In: Clinical physiology of acid-base and electrolyte disorders. 5th ed. New York: McGraw-Hill, 2001: 794–821.

12. Marino PL, Krasner J, O'Moore P. Fluid and electrolyte expert. Philadelphia: WB Saunders, 1987.

13. Makaryus AN, McFarlane SI. Diabetes insipidus: diagnosis and treatment of a complex disease. Cleve Clin J Med 2006;73:65–71. Bibliographic Links 14. Blevins LS Jr, Wand GS. Diabetes insipidus. Crit Care Med 1992;20:69–79. Bibliographic Links 15. Ghirardello S, Malattia C, Scagnelli P, et al. Current perspective on the pathogenesis of central diabetes insipidus. J Pediatr Endocrinol Metab 2005;18:631–645. Bibliographic Links 16. Geheb MA. Clinical approach to the hyperosmolar patient. Crit Care Clin 1987;3:797–815. Bibliographic Links 17. Garofeanu CG, Weir M, Rosas-Arellano MP, et al. Causes of reversible nephrogenic diabetes insipidus: a systematic review. Am J Kidney Dis 2005;45:626–637. Bibliographic Links 18. Moran SM, Jamison RL. The variable hyponatremic response to hyperglycemia. West J Med 1985;142:49–53. Bibliographic Links 19. Ofran Y, Lavi D, Opher D, et al. Fatal voluntary salt intake resulting in the highest ever documented sodium plasma level in adults (255 mmol L-1): a disorder linked to female gender and psychiatric disorders. J Intern Med 2004;256:525–528. Ovid Full TextBibliographic Links

Hyponatremia 20. Adrogue HJ, Madias NE. Hyponatremia. N Engl J Med 2000;342:1581–1589. Ovid Full TextBibliographic Links 21. Diringer MN, Zazulia AR. Hyponatremia in neurologic patients: consequences and approaches to treatment. Neurologist 2006;12:117–126. Ovid Full TextBibliographic Links P.610 22. Tang WW, Kaptein EM, Feinstein EI, et al. Hyponatremia in hospitalized patients with the acquired immunodeficiency syndrome (AIDS) and the AIDS-related complex. Am J Med 1993;94:169–174.

Bibliographic Links 23. Terzian C, Frye EB, Piotrowski ZH. Admission hyponatremia in the elderly: factors influencing prognosis. J Gen Intern Med 1994;9:89–91. Bibliographic Links 24. Ayus JC, Wheeler JM, Arieff AI. Postoperative hyponatremic encephalopathy in menstruant women. Ann Intern Med 1992;117:891–897. Full TextBibliographic Links 25. Weisberg LS. Pseudohyponatremia: a reappraisal. Am J Med 1989; 86:315–318. Bibliographic Links 26. Maesaka JK, Gupta S, Fishbane S. Cerebral salt-wasting syndrome: does it exist? Nephron 1999;82:100–109.

27. Ayus JC, Arieff AI. Pathogenesis and prevention of hyponatremic encephalopathy. Endocrinol Metab Clin North Am 1993;22:425–446. Bibliographic Links 28. Arieff AI, Ayus JC. Pathogenesis of hyponatremic encephalopathy: current concepts. Chest 1993;103:607–610. Bibliographic Links 29. Ayus JC, Arieff AI. Pulmonary complications of hyponatremic encephalopathy: noncardiogenic pulmonary edema and hypercapnic respiratory failure. Chest 1995;107:517–521. Bibliographic Links 30. Brunner JE, Redmond JM, Haggar AM, et al. Central pontine myelinolysis and pontine lesions after rapid correction of hyponatremia: a prospective magnetic resonance imaging study. Ann Neurol 1990;27:61–66. Bibliographic Links

Chapter 33 Potassium Early sea-living organisms exhibited a preference for intracellular potassium and a disdain for intracellular sodium, which eventually changed the composition of the oceans from a potassium salt solution to a sodium salt solution. This behavior is also found in mammalian organisms, in whom potassium is the major intracellular cation and sodium is the major extracellular cation. This pattern is the result of the sodium–potassium exchange pump on cell membranes, which sequesters potassium and extrudes sodium. In humans, only 2% of the total body potassium stores are found outside cells. This lack of extracellular representation limits the value of the plasma (extracellular) potassium concentration as an index of total body potassium stores.

Potassium Distribution The marked discrepancy between the intracellular and extracellular content of potassium is illustrated in Figure 33.1. The total body potassium content in healthy adults is approximately 50 mEq/kg (1,2,3), so a 70-kg adult will have 3500 mEq of total body potassium. However, only 70 mEq (2% of the total amount) is found in the extracellular fluids. Because the plasma accounts for approximately 20% of the extracellular fluid volume, the potassium content of plasma will be about 15 mEq, which is about 0.4% of the total amount of potassium in the body. This suggests that the plasma potassium will be an insensitive marker of changes in total body potassium stores.

Serum Potassium The relationship between changes in total body potassium and changes in serum potassium is curvilinear, as shown in Figure 33.2 (4,5). The slope of the curve decreases on the “deficit” side of the graph, indicating that the change in serum potassium is much smaller when potassium is depleted than when potassium accumulates. In an averaged-size adult with a normal serum potassium concentration (i.e., 3.5 to 5.5 mEq/L), a total body potassium deficit of 200 to 400 mEq is required to produce a 1 mEq/L decrease in serum potassium, whereas a total body potassium P.612 excess of 100 to 200 mEq is required to produce a 1 mEq/L rise in serum potassium ( 5). In other words, potassium depletion must be twice as great as potassium accumulation to produce a significant (1 mEq/L) change in the serum potassium concentration. This difference is due to the large pool of intracellular potassium, which can replenish extracellular stores when potassium is lost.

Figure 33.1 The intracellular and extracellular potassium content in a 70-kg adult with a total body potassium of 50 mEq/L.

View Figure

Hypokalemia Hypokalemia is a serum potassium concentration below 3.5 mEq/L. The causes of hypokalemia can be classified according to whether an intracellular shift of potassium (transcellular shift) occurred or whether a decrease in total body potassium content (potassium depletion) occurred (3,6). The following are some of the possible causes of hypokalemia that are likely to be encountered in the ICU.

Transcellular Shift Potassium movement into cells is facilitated by stimulation of b 2-adrenergic receptors on muscle cell membranes. Inhaled b-agonist bronchodilators (e.g., albuterol) are well known for their ability to reduce the serum potassium concentration, but this effect is mild (0.5 mEq/L or less) in the usual therapeutic doses (7). A more significant effect is seen when inhaled ß-agonists are given in combination with glucose and insulin (7) or diuretics (8). Other factors that promote the transcellular shift of potassium into cells include alkalosis (respiratory or metabolic), hypothermia (accidental or induced), and insulin. Alkalosis has a variable and unpredictable effect on the serum potassium (9). Hypothermia causes a transient drop in serum potassium that usually resolves during P.613 rewarming (10). Lethal cases of hypothermia can be accompanied by hyperkalemia because of widespread cell death (11).

Figure 33.2 The relationship between the serum potassium concentration and changes in total body potassium content. (Redrawn from Brown RS. Extrarenal potassium homeostasis Kidney Int 1986;30:116–127.)

View Figure

Potassium Depletion Potassium depletion can be the result of either renal or extrarenal potassium losses. The site of potassium loss can often be identified by using a combination of urinary potassium and chloride concentrations, as shown in Figure 33.3.

Renal Potassium Loss The leading cause of renal potassium wasting is diuretic therapy. Other causes likely to be seen in the ICU include nasogastric drainage, alkalosis, and magnesium depletion. The urinary chloride is low (less than 15 mEq/L) when nasogastric drainage or alkalosis is involved, and it is high (greater than 25 mEq/L) when magnesium depletion or diuretics are responsible. Magnesium depletion impairs potassium reabsorption across the renal tubules and may play a very important role in promoting and sustaining potassium depletion in critically ill patients, particularly those receiving diuretics (12).

Extrarenal Potassium Loss The major cause of extrarenal potassium loss is diarrhea. The potassium concentration in stool is 75 mEq/L, but because the stool volume is normally 200 mL or less each day, little potassium is lost. In diarrheal states, the daily volume of stool can be as high as 10 L, and thus severe or prolonged diarrhea can result in significant potassium depletion.

Figure 33.3 Diagnostic approach to hypokalemia.

View Figure


Clinical Manifestations Severe hypokalemia (serum K+ below 2.5 mEq/L) can be accompanied by diffuse muscle weakness (3). Milder degrees of hypokalemia (serum K + 2.5 to 3.5 mEq/L) are often asymptomatic. Abnormalities in the ECG, including prominent U waves (more than 1 mm in height), flattening and inversion of T waves, and prolongation of the QT interval, can be present in more than half of the cases (13). None of these changes are specific for hypokalemia. The T wave changes and U waves can be seen with digitalis or left ventricular hypertrophy, and QT prolongation can be seen with hypocalcemia and hypomagnesemia.

Arrhythmias There is a misconception about the ability of hypokalemia to promote cardiac arrhythmias. Hypokalemia alone does not produce serious ventricular cardiac arrhythmias (3,13). Hypokalemia is often combined with other conditions that can promote arrhythmias (e.g., magnesium depletion, digitalis, myocardial ischemia), and the hypokalemia may enhance the proarrhythmic effects of these other conditions (3). P.615 Hypokalemia is well known for its ability to promote digitalis-induced arrhythmias.

Management of Hypokalemia The first concern in hypokalemia is to eliminate or treat any condition that promotes transcellular potassium shifts (e.g., alkalosis) (3). If the hypokalemia is due to potassium

depletion, proceed as described in the following section.

Potassium Deficit If the hypokalemia is due to potassium depletion, a potassium deficit of 10% of the total body potassium stores is expected for every 1 mEq/L decrease in the serum potassium (14). The correlation between potassium deficits and the severity of hypokalemia is shown in Table 33.1. These estimates do not consider any contribution from transcellular potassium shifts, and thus they are meant only as rough guidelines for gauging the severity of potassium depletion.

Potassium Replacement Solutions The usual replacement fluid is potassium chloride, which is available as a concentrated solution (from 1 and 2 mEq/mL) in ampules containing 10, 20, 30, and 40 mEq of potassium. These solutions are extremely hyperosmotic (the 2 mEq/L solution has an osmolality of 4000 mOsm/L H2O) and must be diluted (15). A potassium phosphate solution is also available (contains 4.5 mEq potassium and 3 µM phosphate per mL) and is preferred by some for potassium replacement in diabetic ketoacidosis (because of the phosphate depletion that accompanies ketoacidosis). TABLE 33.1 Potassium Deficits in Hypokalemia*

Serum Potassium (mEq/L)

Potassium Deficit mEq

% Total Body K
















*Estimated deficits for a 70 kg adult with a total body potassium content of 50 mEq/kg.


Infusion Rate The standard method of intravenous potassium replacement is to add 20 mEq of potassium to 100 mL of isotonic saline and infuse this mixture over 1 hour (16). The maximum rate of intravenous potassium replacement is usually set at 20 mEq/hour (16), but dose rates up to 40 mEq/hour occasionally may be necessary (e.g., with serum K + below 1.5 mEq/L or serious arrhythmias), and dose rates as high as 100 mEq/hour have been used safely (17). A large central vein should be used for infusion because of the irritating properties of the hyperosmotic potassium solutions. However, if the desired replacement rate is greater than 20 mEq/hour, the infusion should not be given through a central venous catheter because of the theoretical risk of transient hyperkalemia in the right heart chambers, which can predispose to cardiac standstill. In this situation, the potassium dose can be split and administered via two peripheral veins.

Response The serum potassium may be slow to rise at first, because of the position on the flat part of the curve in Figure 33.2. Full replacement usually takes a few days, particularly if potassium losses are ongoing. If the hypokalemia seems refractory to replacement therapy, the serum magnesium level should be checked. Magnesium depletion promotes urinary potassium losses and can cause refractory hypokalemia (18). The management of hypomagnesemia is presented in Chapter 34.

Hyperkalemia While hypokalemia is often well tolerated, hyperkalemia (serum K + greater than 5.5 mEq/L) can be a serious and life-threatening condition (3,19,20).

Pseudohyperkalemia Potassium release from traumatic hemolysis during the venipuncture can produce a spurious elevation in serum potassium. This is more common than suspected, and has been reported in 20% of blood samples with an elevated serum potassium (21). Potassium release from muscles distal to a tourniquet can also be a source of spuriously high serum potassium levels (22). Because of the risk of spurious hyperkalemia, an unexpected finding of hyperkalemia in an asymptomatic patient should always prompt a repeat measurement before any diagnostic or therapeutic measures are initiated. Potassium release from cells during clot formation in the specimen tube can also produce pseudohyperkalemia when severe leukocytosis (white blood cell count greater than 50,000/mm3) or thrombocytosis (platelet count greater than 1 million/mm 3) is present. When this condition is suspected, the serum potassium should be measured in an unclotted blood sample. P.617

Urine Potassium Hyperkalemia can be caused by potassium release from cells (transcellular shift) or by impaired renal potassium excretion. If the source of the hyperkalemia is unclear, the urinary potassium concentration can be helpful. A high urine potassium (greater than 30 mEq/L) suggests a transcellular shift, and a low urine potassium (less than 30 mEq/L) indicates impaired renal excretion.

Transcellular Shift Acidosis traditionally has been listed as a cause of hyperkalemia because of the tendency for acidosis to both enhance potassium release from cells and reduce renal potassium excretion. However, hyperkalemia does not always accompany respiratory acidosis (9), and no clear evidence exists that organic acidoses (i.e., lactic acidosis and ketoacidosis) can produce hyperkalemia (9). Although hyperkalemia can accompany acidoses associated with renal failure and renal tubular acidosis, hyperkalemia in these instances may be caused by impaired renal potassium excretion. Rhabdomyolysis can release large amounts of potassium into the extracellular fluid, but if renal function is normal, the extra potassium is promptly cleared by the kidneys. For example, severe exercise can raise the serum potassium to 8 mEq/L, but the hyperkalemia resolves with a half-time of 25 seconds (23). Drugs that can promote hyperkalemia via transcellular potassium shifts include ß-receptor antagonists and digitalis (Table 33.2). Serious hyperkalemia (i.e., serum

potassium above 7 mEq/L) is possible only with digitalis toxicity.

Impaired Renal Excretion Renal insufficiency can produce hyperkalemia when the glomerular filtration rate falls below 10 mL/minute or the urine output falls below 1 L/day (24). Exceptions are interstitial nephritis and hyporeninemic P.618 hypoaldosteronism (24). The latter condition is seen in elderly diabetic patients who have defective renin release in response to reduced renal blood flow. TABLE 33.2 Drugs That Can Cause Hyperkalemia

ACE Inhibitors*


Angiotensin Recepter Blockers*



Potassium penicillin





Diuretics (K-sparing)


Heparin ACE = angiotensin converting enzyme, NSAIDs = nonsteroidal antiinflammatory drugs, TMP–SMX = trimethoprim–sulfamethoxazole. *Especially when combined with K-sparing diuretics.

Adrenal insufficiency is a well known cause of hyperkalemia from impaired renal potassium excretion, but is not a common cause of hyperkalemia in the ICU. Drugs that impair renal potassium excretion are considered one of the leading causes of hyperkalemia (3,25). A list of common offenders is shown in Table 33.2. The drugs most commonly implicated are angiotensin-converting enzyme inhibitors, angiotensin receptor blockers, potassium sparing diuretics, and nonsteroidal antiinflammatory drugs (25,26). Other potential offenders in the ICU are heparin, trimethoprim–sulfamethoxazole, and pentamidine (27–29). All of these agents promote hyperkalemia by inhibiting or blocking

the renin–angiotensin–aldosterone system, and all promote hyperkalemia particularly when given with potassium supplements.

Blood Transfusions Massive blood transfusions (i.e., when the transfusion volume exceeds the estimated blood volume) can promote hyperkalemia when given to patients with circulatory shock (30). Potassium leakage from erythrocytes results in a steady rise in plasma potassium levels in stored blood. In whole blood, the plasma potassium rises an average of 1 mEq/L/day. However, because one unit of whole blood contains 250 mL of plasma, this represents an increase of only 0.25 mEq/day in the plasma potassium content per unit of whole blood. After 14 days of storage, the plasma potassium load is 4.4 mEq per unit of whole blood and 3.1 mEq per unit of packed red cells (31). The potassium load in blood transfusions normally is cleared by the kidneys, and thus no sustained rise in plasma potassium occurs. However, in patients with circulatory shock, the extra potassium from blood transfusions can accumulate and produce hyperkalemia. Furthermore, when the volume of distribution for potassium is curtailed by widespread hypoperfusion, the potassium accumulation can be rapid and life-threatening.

Clinical Manifestations The most serious consequence of hyperkalemia is the slowing of electrical conduction in the heart. The ECG can begin to change when the serum potassium reaches 6.0 mEq/L, and it is always abnormal when the serum potassium reaches 8.0 mEq/L (24). Figure 33.4 illustrates the ECG changes associated with progressive hyperkalemia. The earliest change in the ECG is a tall, tapering (tented) T wave that is most evident in precordial leads V2 and V3. Similar “peaked T” waves have been observed in metabolic acidosis (32). As the hyperkalemia progresses, the P wave amplitude decreases and the PR interval lengthens. The P waves eventually disappear and the QRS duration becomes prolonged. The final event is ventricular asystole.

Figure 33.4 The ECG manifestations of progressive hyperkalemia. (Adapted from Burch GE, Winsor T. A primer of electrocardiography. Philadelphia: Lea & Febiger, 1966:143.)

View Figure


Management of Hyperkalemia The acute management of hyperkalemia is guided by the serum potassium level and the ECG (3,20). The therapeutic maneuvers are outlined in Table 33.3.

Membrane Antagonism Calcium directly antagonizes the membrane actions of potassium (33). When hyperkalemia is severe (i.e., above 7 mEq/L) or accompanied by advanced ECG changes (i.e., loss of P waves and prolonged QRS duration), calcium gluconate is administered in the dose shown in Table 33.3. If there P.620 is no response to calcium within a few minutes, a second dose can be given. A third dose will not be effective if there was no response to the second dose of calcium. The response to calcium lasts only 20 or 30 minutes, so other therapies should be initiated to enhance potassium clearance. TABLE 33.3 Acute Management of Hyperkalemia




ECG changes or serum K >7 mEq/L

Calcium gluconate (10%): 10 mL IV over 3 minutes; can repeat in 5 minutes

Response lasts only 20 to 30 minutes. Do not give bicarbonate after calcium.

ECG changes and circulatory compromise

Calcium chloride (10%): 10 mL IV over 3 minutes

Calcium chloride contains 3 times more calcium than calcium gluconate.

AV block refractory to calcium treatment

1. 10 U regular insulin in 500 mL of 20% dextrose: infuse over 1 hour 2. Transvenous pacemaker

Insulin–dextrose treatment should drop the serum K by 1 mEq/L for 1 to 2 hours.

Digitalis cardiotoxicity

1. Magnesium sulfate: 2 g as IV bolus 2. Digitalis specific antibodies if necessary

Do not use calcium for the hyperkalemia of digitalis toxicity.

After acute phase or when no ECG changes

Kayexalate: oral dose of 30 g in 50 mL of 20% sorbitol, or rectal dose of 50 g in 200 mL 20% sorbitol as a retention enema

Oral dosing is preferred. Enemas poorly tolerated by patients and nurses.

Calcium must be given cautiously to patients on digitalis because hypercalcemia can potentiate digitalis cardiotoxicity. For patients receiving digitalis, the calcium gluconate should be added to 100 mL of isotonic saline and infused over 20 to 30 minutes. If the hyperkalemia is a manifestation of digitalis toxicity, calcium is contraindicated. When hyperkalemia is accompanied by evidence of circulatory compromise, calcium chloride is preferred to calcium gluconate. One ampule (10 mL) of 10% calcium chloride contains three times more elemental calcium than one ampule of 10% calcium gluconate (see Table 35.3), and the extra calcium in calcium chloride may prove beneficial in promoting cardiac contraction and maintaining peripheral vascular tone.

Transcellular Shift

Insulin–Dextrose Combined therapy with insulin and dextrose will drive potassium into muscle cells and decrease the serum potassium by an average of 1 mEq/L. P.621 However, this is a temporary effect, and other maneuvers aimed at enhancing potassium clearance are also required.

Sodium Bicarbonate The administration of sodium bicarbonate (44 to 88 mEq) can also shift potassium into cells. However, the most common acidotic condition associated with hyperkalemia is renal failure, and in this condition, insulin–dextrose is much more effective in lowering the serum potassium than bicarbonate (34). Furthermore, bicarbonate binds calcium and should not be given after calcium is administered. For these reasons, there is little value in using bicarbonate to treat hyperkalemia.

Enhanced Clearance Measures aimed at enhancing the removal of potassium from the body can be used alone in mild cases of hyperkalemia (i.e., serum K less than 7 mEq/L) without advanced ECG changes or can serve as a follow-up to calcium and insulin–dextrose therapy.

Exchange Resin Sodium polystyrene sulfonate (Kayexalate) is a cation exchange resin that can enhance potassium clearance across the gastrointestinal mucosa (gastrointestinal dialysis). This resin can be given orally or by retention enema, and it is mixed with 20% sorbitol to prevent concretion. For each mEq of potassium removed, 2 to 3 mEq of sodium are added. If there is concern about the added sodium, one or two doses of furosemide can be used to enhance natriuresis.

Loop Diuretics The loop diuretics furosemide and ethacrynic acid enhance urinary potassium excretion. These agents can be used as a follow-up measure to calcium and insulin–dextrose. This approach is ineffective in renal failure.

Hemodialysis Hemodialysis is the most effective method of lowering the serum potassium in patients with renal failure (3,20).

A Final Word The following points about potassium deserve emphasis: 1. Since only 2% of the potassium is outside cells, it is unlikely that the serum

potassium concentration is an accurate reflection of total body potassium stores. However, we use the serum potassium as a reflection of total body potassium, so there's P.622 a fundamental problem in the way we interpret and manage changes in serum potassium. 2. There is often a rush to correct even mild cases of hypokalemia (serum K + between 3 and 3.5 mEq/L). This is usually not necessary, because hypokalemia is well-tolerated, and does not create a risk of arrhythmias unless there are other arrhythmogenic conditions (such as digitalis toxicity). 3. If hypokalemia is really due to potassium depletion, don't expect an extra 40 mEq of potassium to correct the problem because for each 0.5 mEq/L decrease in serum K+ , you will have to replace about 175 mEq of potassium to replenish total body K+ stores. 4. Don't forget that hypokalemia associated with diuretic therapy is often the result of magnesium depletion, and that potassium replacement will not correct the problem unless magnesium is also replaced.

References Potassium Distribution 1. Rose BD, Post TW. Potassium homeostasis. In: Clinical physiology of acid-base and electrolyte disorders. 5th ed. New York: McGraw-Hill, 2001:372–402.

2. Halperin ML, Kamel KS. Potassium. Lancet 1998;352:135–140. Full TextBibliographic Links 3. Schaefer TJ, Wolford RW. Disorders of potassium. Emerg Med Clin North Am 2005;23:723–747, viii–ix. Bibliographic Links 4. Brown RS. Extrarenal potassium homeostasis. Kidney Int 1986;30:116–127. Full TextBibliographic Links 5. Sterns RH, Cox M, Feig PU, et al. Internal potassium balance and the control of the plasma potassium concentration. Medicine 1981;60:339–354. Bibliographic Links


6. Glover P. Hypokalaemia. Crit Care Resusc 1999;1:239–251. Bibliographic Links 7. Allon M, Copkney C. Albuterol and insulin for treatment of hyperkalemia in hemodialysis patients. Kidney Int 1990;38:869–872. Full TextBibliographic Links 8. Lipworth BJ, McDevitt DG, Struthers AD. Prior treatment with diuretic augments the hypokalemic and electrocardiographic effects of inhaled albuterol. Am J Med 1989;86:653–657. Bibliographic Links 9. Adrogue HJ, Madias NE. Changes in plasma potassium concentration during acute acid-base disturbances. Am J Med 1981;71:456–467. Bibliographic Links 10. Bernard SA, Buist M. Induced hypothermia in critical care medicine: a review. Crit Care Med 2003;31:2041–2051. Ovid Full TextBibliographic Links 11. Schaller MD, Fischer AP, Perret CH. Hyperkalemia: a prognostic factor during acute severe hypothermia. JAMA 1990;264:1842–1845. Bibliographic Links 12. Salem M, Munoz R, Chernow B. Hypomagnesemia in critical illness: a common and clinically important problem. Crit Care Clin 1991;7:225–252. Bibliographic Links 13. Flakeb G, Villarread D, Chapman D. Is hypokalemia a cause of ventricular arrhythmias? J Crit Illness 1986;1:66–74. P.623 14. Stanaszek WF, Romankiewicz JA. Current approaches to management of potassium deficiency. Drug Intell Clin Pharm 1985;19:176–184. Bibliographic Links 15. Trissel LA. Handbook on injectable drugs. 13th ed. Bethesda, MD: American Society of Health System Pharmacists, 2005:1230.

16. Kruse JA, Carlson RW. Rapid correction of hypokalemia using concentrated intravenous potassium chloride infusions. Arch Intern Med 1990;150:613–617. Bibliographic Links

17. Kim GH, Han JS. Therapeutic approach to hypokalemia. Nephron 2002;92 (suppl 1):28–32. Bibliographic Links 18. Whang R, Flink EB, Dyckner T, et al. Magnesium depletion as a cause of refractory potassium repletion. Arch Intern Med 1985;145:1686–1689. Bibliographic Links

Hyperkalemia 19. Williams ME. Endocrine crises: hyperkalemia. Crit Care Clin 1991;7:155–174. Bibliographic Links 20. Evans KJ, Greenberg A. Hyperkalemia: a review. J Intensive Care Med 2005; 20:272–290. Bibliographic Links 21. Rimmer JM, Horn JF, Gennari FJ. Hyperkalemia as a complication of drug therapy. Arch Intern Med 1987;147:867–869. Bibliographic Links 22. Don BR, Sebastian A, Cheitlin M, et al. Pseudohyperkalemia caused by fist clenching during phlebotomy. N Engl J Med 1990;322:1290–1292. Bibliographic Links 23. Medbo JI, Sejersted OM. Plasma potassium changes with high intensity exercise. J Physiol 1990;421:105–122. Bibliographic Links 24. Williams ME, Rosa RM. Hyperkalemia: disorders of internal and external potassium balance. J Intensive Care Med 1988;3:52–64.

25. Perazella MA. Drug-induced hyperkalemia: old culprits and new offenders. Am J Med 2000;109:307–314. Bibliographic Links 26. Palmer BF. Managing hyperkalemia caused by inhibitors of the renin-angiotensin-aldosterone system. N Engl J Med 2004;351:585–592. Ovid Full TextBibliographic Links 27. Oster JR, Singer I, Fishman LM. Heparin-induced aldosterone suppression and hyperkalemia. Am J Med 1995;98:575–586.

Bibliographic Links 28. Greenberg S, Reiser IW, Chou SY, et al. Trimethoprim-sulfamethoxazole induces reversible hyperkalemia. Ann Intern Med 1993;119:291–295.Ovid Full TextFull TextBibliographic Links 29. Peltz S, Hashmi S. Pentamidine-induced severe hyperkalemia. Am J Med 1989;87:698–699. Bibliographic Links 30. Leveen HH, Pasternack HS, Lustrin I, et al. Hemorrhage and transfusion as the major cause of cardiac arrest. JAMA 1960;173:770–777. Bibliographic Links 31. Michael JM, Dorner I, Bruns D, et al. Potassium load in CPD-preserved whole blood and two types of packed red blood cells. Transfusion 1975;15:144–149. Bibliographic Links 32. Dreyfuss D, Jondeau G, Couturier R, et al. Tall T waves during metabolic acidosis without hyperkalemia: a prospective study. Crit Care Med 1989;17: 404–408. Bibliographic Links 33. Bosnjak ZJ, Lynch C, III. Cardiac electrophysiology. In: Yaksh T, Lynch C, III, Zapol WM, et al, eds. Anesthesia: biologic foundations. New York: Lippincott-Raven, 1998:1001–1040.

34. Blumberg A, Weidmann P, Shaw S, et al. Effect of various therapeutic approaches on plasma potassium and major regulating factors in terminal renal failure. Am J Med 1988;85:507–512. Bibliographic Links

Chapter 34 Magnesium Magnesium is the second most abundant intracellular cation in the human body (potassium being the first), where it serves as a cofactor for more than 300 enzyme reactions that involve adenosine triphosphate (1,2,3,4). One of the magnesium-dependent enzyme systems is the membrane pump that generates the electrical gradient across cell membranes. As a result, magnesium plays an important role in the activity of electrically excitable tissues (1,2,5,6,7). Magnesium also regulates the movement of calcium into smooth muscle cells, which gives it a pivotal role in the maintenance of cardiac contractile strength and peripheral vascular tone ( 5).

Magnesium Balance The content and distribution of magnesium in the human body is shown in Table 34.1 (8). The average-size adult contains approximately 24 g (1 mole, or 2000 mEq) of magnesium; a little over half is located in bone, whereas less than 1% is located in plasma. This lack of representation in the plasma limits the value of the plasma magnesium concentration as an index of total body magnesium stores. This is particularly true in patients with magnesium deficiency, in whom serum magnesium levels can be normal in the face of total body magnesium depletion (8,9).

Serum Magnesium Serum is favored over plasma for magnesium assays because the anticoagulant used for plasma samples can be contaminated with citrate or other anions that bind magnesium (8). The normal range for serum magnesium depends on the daily magnesium intake, which varies according to geographic region. The normal range for healthy adults residing in the United States is shown in Table 34.2 (10).

Ionized Magnesium Only 67% of the magnesium in plasma is in the ionized (active) form, and the remaining 33% is either bound to plasma proteins (19% of the total) or chelated with divalent anions such as phosphate and sulfate (14% of the total) (11). The standard assay for magnesium (i.e., spectrophotometry) P.626 measures all three fractions of magnesium. Therefore, when the serum magnesium is abnormally low, it is impossible to determine whether the problem is a decrease in the ionized (active) fraction or a decrease in the bound fractions (e.g., hypoproteinemia) ( 12). The level of ionized magnesium can be measured with an ion-specific electrode ( 13) or

by ultrafiltration of plasma (14), but these techniques are not routinely available for clinical use. However, because the total amount of magnesium in plasma is small, the difference between the ionized and bound magnesium content may not be large enough to be clinically relevant. TABLE 34.1 Magnesium Distribution in Adults


Weight (kg)

Magnesium Content (mEq)

Total Body Magnesium (%)









Soft Tissue













70 kg

2000 mEq


From: Elin RJ. Assessment of magnesium status. Clin Chem 1987;33:1965–1970.

Urinary Magnesium The normal range for urinary magnesium excretion is shown in Table 34.2. Under normal circumstances, only small quantities of magnesium are excreted in the urine. When magnesium intake is deficient, the kidneys conserve magnesium and urinary magnesium excretion falls to negligible levels. This is shown in Figure 34.1. After the start of a magnesium deficient diet, the urinary magnesium excretion promptly falls to negligible levels and the serum magnesium remains in the normal range. This illustrates the relative value of urinary magnesium over serum magnesium levels in the detection of magnesium deficiency. This is discussed again later in this chapter. TABLE 34.2 Reference Ranges for Magnesium*


Traditional Units

SI Units

Serum magnesium: Total

1.4–2.0 mEq/L

0.7–1.0 mmol/L


0.8–1.1 mEq/L

0.4–0.6 mmol/L

5–15 mEq/24 hr

2.5–7.5 mmol/24 hr

Urinary magnesium

*Pertains to healthy adults residing in the United States. From: Lowenstein FW, Stanton MF. Serum magnesium levels in the United States, 1971–1974. J Am Coll Nutr 1986;5:399–414.

View Figure

Figure 34.1 Urinary and plasma magnesium levels in a healthy volunteer placed on a magnesium-free diet. Solid bars on the vertical axes indicate the normal range for urine and plasma magnesium. (Adapted from Shils ME. Experimental human magnesium deficiency. Medicine 1969;48:61–82.Bibliographic Links)


Magnesium Deficiency Magnesium deficiency is common in hospitalized patients (1,2,3). Hypo-magnesemia is reported in up to 20% of patients on medical wards and in as many as 65% of patients in ICUs (1,2,3). Because magnesium depletion may not be accompanied by hypomagnesemia, the incidence of magnesium depletion is even higher than indicated by the surveys of hypomagnesemia. In fact, magnesium depletion has been described as “the most underdiagnosed electrolyte abnormality in current medical practice” ( 15).

Predisposing Conditions Because serum magnesium levels have a limited ability to detect magnesium depletion, recognizing the conditions that predispose to magnesium depletion may be the only clue of an underlying electrolyte imbalance. The common predisposing conditions for magnesium depletion are listed in Table 34.3.

Diuretic Therapy Diuretics are the leading cause of magnesium deficiency. Drug-induced inhibition of sodium reabsorption also interferes with magnesium reabsorption, and the resultant urinary magnesium losses can parallel urinary sodium losses. Urinary magnesium excretion is most pronounced with the loop diuretics (furosemide and ethacrynic acid). Magnesium deficiency has been reported in 50% of patients receiving chronic therapy P.628 with furosemide (16). The thiazide diuretics show a similar tendency for magnesium depletion, but only in elderly patients (17). Magnesium depletion is not a complication of therapy with “potassium-sparing” diuretics such as triamterene ( 18). TABLE 34.3 Markers of Possible Magnesium Depletion

Predisposing Conditions Drug therapy:*

Clinical Findings Electrolyte abnormalities:*

Furosemide (50%)

Hypokalemia (40%)

Aminoglycosides (30%)

Hypophosphatemia (30%)

Amphotericin, pentamidine

Hyponatremia (27%)

Digitalis (20%)

Hypocalcemia (22%)

Cisplatin, cyclosporine

Cardiac manifestations:

Diarrhea (secretory)


Alcohol abuse (chronic)

Arrhythmias (refractory)

Diabetes mellitus

Digitalis toxicity

Acute MI

Hyperactive CNS Syndrome

*Numbers in parentheses indicate incidence of associated hypomagnesemia.

Antibiotic Therapy The antibiotics that promote magnesium depletion are the aminoglycosides, amphotericin and pentamidine (19,20). The aminoglycosides block magnesium reabsorption in the ascending loop of Henle, and hypomagnesemia has been reported in 30% of patients receiving aminoglycoside therapy (20). The other risk associated with antibiotic use occurs with antibiotic-associated diarrhea, which can be accompanied by significant magnesium losses in the stool.

Other Drugs A variety of other drugs have been associated with magnesium depletion, including digitalis, epinephrine, and the chemotherapeutic agents cisplatin and cyclosporine (19,21). The first two agents shift magnesium into cells, whereas the latter two promote renal magnesium excretion.

Alcohol-Related Illness Hypomagnesemia is reported in 30% of hospital admissions for alcohol abuse, and in 85% of admissions for delirium tremens (22,23). The magnesium depletion in these conditions is due to a number of factors, including generalized malnutrition and chronic diarrhea. In addition, there is an association between magnesium deficiency and thiamine deficiency (24). Magnesium is required for the transformation of thiamine into thiamine pyrophosphate, so magnesium deficiency can promote thiamine deficiency in the face of adequate thiamine intake. For P.629 this reason, the magnesium status should be monitored periodically in patients receiving daily thiamine supplements.

Secretory Diarrhea A high concentration of magnesium (10 to 14 mEq/L) is present in secretions from the lower gastrointestinal tract (25), and thus secretory diarrhea can be accompanied by profound magnesium depletion (23). Upper tract secretions are not rich in magnesium (1 to 2 mEq/L), so vomiting does not pose a risk for magnesium depletion.

Diabetes Mellitus Magnesium depletion is common in insulin-dependent diabetic patients, probably as a result of urinary magnesium losses that accompany glycosuria (26). Hypomagnesemia is reported in only 7% of admissions for diabetic ketoacidosis, but the incidence increases to 50% over the first 12 hours after admission (27), probably as a result of insulin-induced movement of magnesium into cells.

Acute Myocardial Infarction As many as 80% of patients with acute myocardial infarction (MI) can have hypomagnesemia in the first 48 hours after the event (28). The mechanism is unclear but may be due to an intracellular shift of magnesium caused by endogenous catecholamine excess.

Clinical Manifestations Although no clinical manifestations are specific for magnesium deficiency, the following clinical findings are suggestive of an underlying magnesium deficiency ( Table 34.3).

Associated Electrolyte Abnormalities Magnesium depletion is often accompanied by depletion of other electrolytes, such as potassium, phosphate, and calcium (See Table 34.3) (29). As mentioned in Chapter 33, the hypokalemia that accompanies magnesium depletion can be refractory to potassium replacement therapy, and magnesium repletion is often necessary before potassium repletion is possible (30). The hypocalcemia that accompanies magnesium depletion is due to impaired parathormone release (31) combined with an impaired end-organ response to parathormone (32). In addition, magnesium deficiency may act on bone directly to reduce calcium release, independent of parathyroid hormone (33). As with the hypokalemia, the hypocalcemia from magnesium depletion is difficult to correct unless magnesium deficits are corrected. Hypophosphatemia is a cause rather than effect of magnesium depletion. The mechanism is enhanced renal magnesium excretion (34). Therefore, when hypophosphatemia accompanies hypomagnesemia, the P.630 phosphate stores should be replenished to ensure adequate repletion of magnesium stores.

Arrhythmias Because magnesium is required for proper function of the membrane pump on cardiac cell membranes, magnesium depletion will depolarize cardiac cells and promote tachyarrhythmias. Because both digitalis and magnesium deficiency act to inhibit the membrane pump, magnesium deficiency will magnify the digitalis effect and promote digitalis cardiotoxicity. Intravenous magnesium can suppress digitalis-toxic arrhythmias, even when serum magnesium levels are normal (35,36). Intravenous magnesium can also abolish refractory arrhythmias (i.e., unresponsive to traditional antiarrhythmic agents) in the absence of hypomagnesemia (37). This effect may be due to a membrane-stabilizing effect of magnesium that is unrelated to magnesium repletion. One of the most serious arrhythmias associated with magnesium depletion is torsades de pointes (polymorphous ventricular tachycardia). The role of magnesium in this arrhythmia is discussed in Chapter 18.

Neurologic Findings The neurologic manifestations of magnesium deficiency include altered mentation,

generalized seizures, tremors, and hyperreflexia. All are uncommon, nonspecific, and have little diagnostic value. A neurologic syndrome described recently that can abate with magnesium therapy deserves mention. The clinical presentation is characterized by ataxia, slurred speech, metabolic acidosis, excessive salivation, diffuse muscle spasms, generalized seizures, and progressive obtundation (38). The clinical features are often brought out by loud noises or bodily contact, and thus the term reactive central nervous system magnesium deficiency has been used to describe this disorder. This syndrome is associated with reduced magnesium levels in cerebrospinal fluid, and it resolves with magnesium infusion. The prevalence of this disorder is unknown at present.

Diagnosis As mentioned several times, the serum magnesium level is an insensitive marker of magnesium depletion. When magnesium depletion is due to nonrenal factors (e.g., diarrhea), the urinary magnesium excretion is a more sensitive test for magnesium depletion (39). However, because most cases of magnesium depletion are due to enhanced renal magnesium excretion, the diagnostic value of urinary magnesium excretion may be limited.

Magnesium Retention Test In the absence of renal magnesium wasting, the urinary excretion of magnesium in response to a magnesium load may be the most sensitive index of total body magnesium stores (40,41). This method is outlined P.631 in Table 34.4. The normal rate of magnesium reabsorption is close to the maximum tubular reabsorption rate (Tmax), so most of the infused magnesium will be excreted in the urine when magnesium stores are normal. However, when magnesium stores are deficient, the magnesium reabsorption rate is much lower than the T max, so more of the infused magnesium will be reabsorbed and less will be excreted in the urine. When less than 50% of the infused magnesium is recovered in the urine, magnesium deficiency is likely, and when more than 80% of the infused magnesium is excreted in the urine, magnesium deficiency is unlikely. This test can be particularly valuable in determining the end-point of magnesium replacement therapy (i.e., magnesium replacement is continued until urinary magnesium excretion is at least 80% of the infused magnesium load). It is important to emphasize that this test will be unreliable in patients with impaired renal function or when there is ongoing renal magnesium wasting. TABLE 34.4 Renal Magnesium Retention Test

Indications: 1. For suspected magnesium deficiency when the serum magnesium concentration is normal. 2. Can be useful for determining the end-point of magnesium replacement therapy. 3. Is not reliable in the setting of renal magnesium wasting or when renal function is impaired. Contraindications: 1. Cardiovascular instability or renal failure. Methodology*: 1. Add 24 mmol of magnesium (6 g of MgSO4) to 250 mL of isotonic saline and infuse over 1 hour. 2. Collect urine for 24 hours, beginning at the onset of the magnesium infusion. 3. A urinary magnesium excretion of less than 12 mmol (24 mEq) in 24 hours (i.e., less than 50% of the infused magnesium) is evidence of total body magnesium depletion. *Magnesium infusion protocol. From: Clague JE, Edwards RH, Jackson MJ. Intravenous magnesium loading in chronic fatigue syndrome. Lancet 1992;340:124–125.

Magnesium Replacement Therapy Preparations The magnesium preparations available for oral and parenteral use are listed in Table 34.5 (42,43). The oral preparations can be used for daily maintenance therapy (5 mg/kg in normal subjects) and for correcting P.632 mild, asymptomatic magnesium deficiency. However, because intestinal absorption of oral magnesium is erratic, parenteral magnesium is preferred for treating symptomatic or severe magnesium deficiency. TABLE 34.5 Oral and Parenteral Magnesium Preparations


Elemental Magnesium

Oral preparations: Magnesium chloride enteric coated tablets

64 mg (5.3 mEq)

Magnesium oxide tablets (400 mg)

241 mg (19.8 mEq)

Magnesium oxide tablets (140 mg)

85 mg (6.9 mEq)

Magnesium gluconate tablets (500 mg)

27 mg (2.3 mEq)

Parenteral solutions: Magnesium sulfate (50%)*

500 mg/mL (4 mEq/mL)

Magnesium sulfate (12.5%)

120 mg/mL (1 mEq/mL)

*Should be diluted to a 20% solution for intravenous injection.

Magnesium Sulfate The standard intravenous preparation is magnesium sulfate (MgSO 4). Each gram of magnesium sulfate has 8 mEq (4 mmol) of elemental magnesium (6). A 50% magnesium sulfate solution (500 mg/mL) has an osmolarity of 4000 mOsm/L (43), so it must be diluted to a 10% (100 mg/mL) or 20% (200 mg/mL) solution for intravenous use. Saline solutions should be used as the diluent for magnesium sulfate. Ringer's solutions should not be used because the calcium in Ringer's solutions will counteract the actions of the infused magnesium.

Replacement Protocols The following magnesium replacement protocols are recommended for patients with normal renal function (44).

Mild, Asymptomatic Hypomagnesemia The following guidelines can be used for patients with mild hypomagnesemia and no

apparent complications (44): 1. Assume a total magnesium deficit of 1 to 2 mEq/kg. 2. Because 50% of the infused magnesium can be lost in the urine, assume that the total magnesium requirement is twice the magnesium deficit. 3. Replace 1 mEq/kg for the first 24 hours, and 0.5 mEq/kg daily for the next 3 to 5 days. 4. If the serum magnesium is greater than 1 mEq/L, oral magnesium can be used for replacement therapy. P.633

Moderate Hypomagnesemia The following therapy is intended for patients with a serum magnesium level less than 1 mEq/L or when hypomagnesemia is accompanied by other electrolyte abnormalities: 1. Add 6 g MgSO 4 (48 mEq Mg) to 250 or 500 mL isotonic saline and infuse over 3 hours. 2. Follow with 5 g MgSO4 (40 mEq Mg) in 250 or 500 mL isotonic saline infused over the next 6 hours. 3. Continue with 5 g MgSO 4 every 12 hours (by continuous infusion) for the next 5 days.

Life-Threatening Hypomagnesemia When hypomagnesemia is associated with serious cardiac arrhythmias (e.g., torsades de pointes) or generalized seizures, do the following: 1. Infuse 2 g MgSO4 (16 mEq Mg) intravenously over 2–5 minutes. 2. Follow with 5 g MgSO4 (40 mEq Mg) in 250 or 500 mL isotonic saline infused over the next 6 hours. 3. Continue with 5 g MgSO 4 every 12 hours (by continuous infusion) for the next 5 days. Serum magnesium levels will rise after the initial magnesium bolus but will begin to fall after 15 minutes. Therefore, it is important to follow the bolus dose with a continuous magnesium infusion. Serum magnesium levels may normalize after 1 to 2 days, but it will take several days to replenish the total body magnesium stores.

Hypomagnesemia and Renal Insufficiency Hypomagnesemia is not common in renal insufficiency but can occur when severe or chronic diarrhea is present and the creatinine clearance is greater than 30 mL/minute. When magnesium is replaced in the setting of renal insufficiency, no more than 50% of

the magnesium in the standard replacement protocols should be administered ( 44), and the serum magnesium should be monitored carefully.

Magnesium Accumulation Magnesium accumulation occurs almost exclusively in patients with impaired renal function. In one survey of hospitalized patients, hypermagnesemia (i.e., a serum magnesium greater than 2 mEq/L) was reported in 5% of patients (45).

Predisposing Conditions Hemolysis The magnesium concentration in erythrocytes is approximately three times greater than that in serum (46), so hemolysis can increase the P.634 plasma magnesium. This can occur either in vivo from a hemolytic anemia or in vitro from traumatic disruption of erythrocytes during phlebotomy. In hemolytic anemia, the serum magnesium is expected to rise by 0.1 mEq/L for every 250 mL of erythrocytes that lyse completely (46), so hypermagnesemia is expected only with massive hemolysis.

Renal Insufficiency The renal excretion of magnesium becomes impaired when the creatinine clearance falls below 30 mL/minute (47). However, hypermagnesemia is not a prominent feature of renal insufficiency unless magnesium intake is increased.

Others Other conditions that can predispose to mild hypermagnesemia are diabetic ketoacidosis (transient), adrenal insufficiency, hyperparathyroidism, and lithium intoxication (47).

Clinical Features The clinical consequences of progressive hypermagnesemia are listed below (47). Manifestation

Serum Magnesium


>4 mEq/L

1st° AV Block

>5 mEq/L

Complete Heart Block

>10 mEq/L

Cardiac Arrest

>13 mEq/L

Magnesium has been described as nature's physiologic calcium blocker (48), and most of the serious consequences of hypermagnesemia are due to calcium antagonism in the cardiovascular system. Most of the cardiovascular depression is the result of cardiac conduction delays. Depressed contractility and vasodilation are not prominent.

Management Hemodialysis is the treatment of choice for severe hypermagnesemia. Intravenous calcium gluconate (1 g IV over 2 to 3 minutes) can be used to antagonize the cardiovascular effects of hypermagnesemia temporarily, until dialysis is started (49). If fluids are permissible and some renal function is preserved, aggressive volume infusion combined with furosemide may be effective in reducing the serum magnesium levels in less advanced cases of hypermagnesemia. P.635

A Final Word The following points about magnesium deserve emphasis: 1. Because 99% of the magnesium in the body is inside cells, the serum magnesium is not a sensitive marker of total body magnesium stores, and serum magnesium levels can be normal in patients who are magnesium depleted. The urine magnesium is a better marker of magnesium depletion (except in patients receiving furosemide, which increases urinary magnesium losses). 2. Magnesium depletion is probably very common in ICU patients, particularly in patients with secretory diarrhea and patients receiving furosemide and aminoglycosides. 3. Magnesium is a cofactor for all ATPase reactions, so magnesium depletion could lead to defects in cellular energy utilization. 4. Magnesium should be given daily to all ICU patients except those with renal insufficiency. Magnesium supplements are particularly important in patients receiving furosemide. 5. Magnesium depletion may be the cause of diuretic-associated hypokalemia, and magnesium repletion is often necessary in these cases before the serum potassium will return to normal. 6. In patients with hypomagnesemia, magnesium replacement will correct the serum magnesium before total body stores of magnesium are repleted. The best indicator of magnesium repletion is the urinary excretion of magnesium (see Table 34.4)

References General Reviews

1. Noronha JL, Matuschak GM. Magnesium in critical illness: metabolism, assessment, and treatment. Intensive Care Med 2002;28:667–679. Bibliographic Links 2. Tong GM, Rude RK. Magnesium deficiency in critical illness. J Intensive Care Med 2005;20:3–17. Bibliographic Links 3. Weisinger JR, Bellorin-Font E. Magnesium and phosphorus. Lancet 1998;352: 391–396. Full TextBibliographic Links 4. Rude RK, Shils ME. Magnesium. In: Shils ME, Shike M, Ross AC, et al. eds. Modern nutrition in health and disease. 10th ed. Philadelphia: Lippincott Williams & Wilkins, 2006:223–247.

5. White RE, Hartzell HC. Magnesium ions in cardiac function: regulator of ion channels and second messengers. Biochem Pharmacol 1989;38:859–867. Full TextBibliographic Links 6. McLean RM. Magnesium and its therapeutic uses: a review. Am J Med 1994; 96:63–76. Bibliographic Links 7. Marino PL. Calcium and magnesium in critical illness: a practical approach. In: Sivak ED, Higgins TL, Seiver A, eds. The high risk patient: management of the critically ill. Baltimore: Williams & Wilkins, 1995:1183–1195. P.636

Magnesium Balance 8. Elin RJ. Assessment of magnesium status. Clin Chem 1987;33:1965–1970. Bibliographic Links 9. Reinhart RA. Magnesium metabolism: a review with special reference to the relationship between intracellular content and serum levels. Arch Intern Med 1988;148:2415–2420. Bibliographic Links 10. Lowenstein FW, Stanton MF. Serum magnesium levels in the United States, 1971–1974. J Am Coll Nutr 1986;5:399–414.

Bibliographic Links 11. Altura BT, Altura BM. A method for distinguishing ionized, complexed and protein-bound Mg in normal and diseased subjects. Scand J Clin Lab Invest 1994;217:83–87. Bibliographic Links 12. Kroll MH, Elin RJ. Relationships between magnesium and protein concentrations in serum. Clin Chem 1985;31:244–246. Bibliographic Links 13. Alvarez-Leefmans FJ, Giraldez F, Gamino SM. Intracellular free magnesium in excitable cells: its measurement and its biologic significance. Can J Physiol Pharmacol 1987;65:915–925. Bibliographic Links 14. Munoz R, Khilnani P, Salem M. Ionized hypomagnesemia: a frequent problem in critically ill neonates. Crit Care Med 1991;19:S48.

Magnesium Depletion 15. Whang R. Magnesium deficiency: pathogenesis, prevalence, and clinical implications. Am J Med 1987;82:24–29. Bibliographic Links 16. Dyckner T, Wester PO. Potassium/magnesium depletion in patients with cardiovascular disease. Am J Med 1987;82:11–17. Bibliographic Links 17. Hollifield JW. Thiazide treatment of systemic hypertension: effects on serum magnesium and ventricular ectopic activity. Am J Cardiol 1989;63:22G–25G. Bibliographic Links 18. Ryan MP. Diuretics and potassium/magnesium depletion: directions for treatment. Am J Med 1987;82:38–47. Bibliographic Links 19. Atsmon J, Dolev E. Drug-induced hypomagnesaemia: scope and management. Drug Safety 2005;28:763–788. Full TextBibliographic Links 20. Zaloga GP, Chernow B, Pock A, et al. Hypomagnesemia is a common complication of aminoglycoside therapy. Surg Gynecol Obstet 1984;158:561–565.

Bibliographic Links 21. Whang R, Oei TO, Watanabe A. Frequency of hypomagnesemia in hospitalized patients receiving digitalis. Arch Intern Med 1985;145:655–656. Bibliographic Links 22. Balesteri FJ. Magnesium metabolism in the critically ill. Crit Care Clin 1985;5: 217–226.

23. Martin HE. Clinical magnesium deficiency. Ann N Y Acad Sci 1969;162:891–900. Bibliographic Links 24. Dyckner T, Ek B, Nyhlin H, et al. Aggravation of thiamine deficiency by magnesium depletion: a case report. Acta Med Scand 1985;218:129–131. Bibliographic Links 25. Kassirer J, Hricik D, Cohen J. Repairing body fluids: principles and practice. 1st ed. Philadelphia: WB Saunders, 1989:118–129.

26. Sjogren A, Floren CH, Nilsson A. Magnesium deficiency in IDDM related to level of glycosylated hemoglobin. Diabetes 1986;35:459–463. Bibliographic Links 27. Lau K. Magnesium metabolism: normal and abnormal. In: Arieff AI, DeFronzo RA, eds. Fluids, electrolytes, and acid base disorders. New York: Churchill Livingstone, 1985:575–623. P.637 28. Abraham AS, Rosenmann D, Kramer M, et al. Magnesium in the prevention of lethal arrhythmias in acute myocardial infarction. Arch Intern Med 1987; 147:753–755. Bibliographic Links

Clinical Manifestations 29. Whang R, Oei TO, Aikawa JK, et al. Predictors of clinical hypomagnesemia: hypokalemia, hypophosphatemia, hyponatremia, and hypocalcemia. Arch Intern Med 1984;144:1794–1796. Bibliographic Links

30. Whang R, Flink EB, Dyckner T, et al. Magnesium depletion as a cause of refractory potassium repletion. Arch Intern Med 1985;145:1686–1689. Bibliographic Links 31. Anast CS, Winnacker JL, Forte LR, et al. Impaired release of parathyroid hormone in magnesium deficiency. J Clin Endocrinol Metab 1976;42:707–717. Bibliographic Links 32. Rude RK, Oldham SB, Singer FR. Functional hypoparathyroidism and parathyroid hormone end-organ resistance in human magnesium deficiency. Clin Endocrinol 1976;5:209–224. Bibliographic Links 33. Graber ML, Schulman G. Hypomagnesemic hypocalcemia independent of parathyroid hormone. Ann Intern Med 1986;104:804–805. Full TextBibliographic Links 34. Dominguez JH, Gray RW, Lemann J Jr. Dietary phosphate deprivation in women and men: effects on mineral and acid balances, parathyroid hormone and the metabolism of 25-OH-vitamin D. J Clin Endocrinol Metab 1976;43: 1056–1068. Bibliographic Links 35. Cohen L, Kitzes R. Magnesium sulfate and digitalis-toxic arrhythmias. JAMA 1983;249:2808–2810. Bibliographic Links 36. French JH, Thomas RG, Siskind AP, et al. Magnesium therapy in massive digoxin intoxication. Ann Emerg Med 1984;13:562–566. Full TextBibliographic Links 37. Tzivoni D, Keren A. Suppression of ventricular arrhythmias by magnesium. Am J Cardiol 1990;65:1397–1399. Bibliographic Links 38. Langley WF, Mann D. Central nervous system magnesium deficiency. Arch Intern Med 1991;151:593–596. Bibliographic Links

Diagnosis 39. Fleming CR, George L, Stoner GL, et al. The importance of urinary magnesium values in patients with gut failure. Mayo Clin Proc 1996;71:21–24. Bibliographic Links

40. Clague JE, Edwards RH, Jackson MJ. Intravenous magnesium loading in chronic fatigue syndrome. Lancet 1992;340:124–125. Full TextBibliographic Links 41. Hebert P, Mehta N, Wang J, et al. Functional magnesium deficiency in critically ill patients identified using a magnesium-loading test. Crit Care Med 1997;25:749–755. Ovid Full TextBibliographic Links

Magnesium Replacement Therapy 42. DiPalma JR. Magnesium replacement therapy. Am Fam Physician 1990;42: 173–176. Bibliographic Links 43. Trissel LA. Handbook on injectable drugs. 13th ed. Bethesda, MD: American Social Health System Pharmacists, 2005.

44. Oster JR, Epstein M. Management of magnesium depletion. Am J Nephrol 1988;8:349–354. Bibliographic Links P.638

Magnesium Accumulation 45. Whang R, Ryder KW. Frequency of hypomagnesemia and hypermagnesemia: requested vs routine. JAMA 1990;263:3063–3064. Bibliographic Links 46. Elin RJ. Magnesium metabolism in health and disease. Dis Mon 1988;34:161–218. Bibliographic Links 47. Van Hook JW. Hypermagnesemia. Crit Care Clin 1991;7:215–223. Bibliographic Links 48. Iseri LT, French JH. Magnesium: nature's physiologic calcium blocker. Am Heart J 1984;108:188–193. Full TextBibliographic Links 49. Mordes JP, Wacker WE. Excess magnesium. Pharmacol Rev 1977;29:273–300

Bibliographic Links

Chapter 35 Calcium and Phosphorus Calcium and phosphorus are responsible for much of the structural integrity of the bony skeleton. Although neither is found in abundance in the soft tissues, both play an important role in vital cell functions. Phosphorus participates in aerobic energy production, whereas calcium participates in several diverse processes, such as blood coagulation, neuromuscular transmission, and smooth muscle contraction. Considering the important functions of these electrolytes, it is surprising that abnormalities in calcium and phosphorus balance are so well tolerated.

Calcium Calcium is the most abundant electrolyte in the human body (the average adult has more than half a kilogram of calcium), but 99% is in bone (1,2). In the soft tissues, calcium is 10,000 times more concentrated than in the extracellular fluids (2,3).

Plasma Calcium The calcium in plasma is present in three forms, as depicted in Figure 35.1. Approximately half of the calcium is ionized (biologically active) and the remainder is complexed (biologically inactive) (1). In the inactive form, 80% of calcium is bound to albumin, while 20% is complexed to plasma anions such as proteins and sulfates. The concentration of total and ionized calcium in plasma is shown in Table 35.1. These values may vary slightly in different clinical laboratories.

Total versus Ionized Calcium The calcium assay used by most clinical laboratories measures all three fractions of calcium, which can be misleading. The column on the right in Figure 35.1 demonstrates the effects of a decrease in the concentration of albumin in plasma. Because albumin is responsible for 80% of the protein-bound calcium in plasma, a decrease in albumin decreases the P.640 amount of calcium in the protein-bound fraction. The total calcium in plasma decreases by the same amount, but the ionized calcium remains unchanged. Because the ionized calcium is the physiologically active fraction, the hypocalcemia caused by hypoalbuminemia is not physiologically significant. The hypocalcemia that is physiologically significant is ionized hypocalcemia (4,5).

Figure 35.1 The three fractions of calcium in plasma and the contribution of each to the total calcium concentration. The column on the right shows how a decrease in plasma albumin can reduce the total plasma calcium without affecting the ionized calcium. View Figure

A variety of correction factors have been proposed for adjusting the plasma calcium concentration in patients with hypoalbuminemia. However, none of these correction factors are reliable (4,6), and the only method of identifying true (ionized) hypocalcemia in the face of hypoalbuminemia is to measure the ionized fraction of calcium in plasma.

Ionized Calcium Measurement Ionized calcium can be measured in whole blood, plasma, or serum with ion-specific electrodes that are now available in most clinical laboratories P.641 (7). The normal concentration of ionized calcium in plasma is shown in Table 35.1. TABLE 35.1 Normal Ranges for Calcium and Phosphorus in Blood

Serum Electrolyte

Traditional Units (mg/dL)

Conversion Factor*

SI Units (mmol/L)

Total calcium




Ionized calcium








* Multiply traditional units by conversion factor to derive SI Units or divide SI units by conversion factor to derive tranditional units.

Blood Collection Several conditions can alter the level of ionized calcium in blood samples (7). Acidosis decreases the binding of calcium to albumin and increases the ionized calcium, whereas alkalosis has the opposite effect. Loss of carbon dioxide from a blood sample could falsely lower the ionized calcium, so it is important to avoid gas bubbles in the blood sample. Anticoagulants (e.g., heparin, citrate, and EDTA) can bind calcium, so blood samples should not be placed in collection tubes that contain these anticoagulants. Tubes with red stoppers (“red top” tubes) contain silicone and are adequate for measuring ionized calcium in serum samples. Heparinized syringes can be used for measuring ionized calcium in whole blood. Although heparin also binds calcium, the effect is minimal if the heparin level is less than 15 U/mL of blood (7).

Ionized Hypocalcemia Ionized hypocalcemia has been reported in 15 to 50% of admissions to the ICU (8). The common disorders associated with ionized hypocalcemia in ICU patients are listed in Table 35.2. Hypoparathyroidism is a leading cause of hypocalcemia in outpatients, but is not a consideration in the ICU unless neck surgery has been performed recently.

Predisposing Conditions Magnesium Depletion Magnesium depletion promotes hypocalcemia by inhibiting parathormone secretion and reducing end-organ responsiveness to parathormone (see Chapter 34). Hypocalcemia from magnesium depletion is refractory to calcium replacement therapy, and magnesium repletion often corrects the hypocalcemia without calcium replacement.

TABLE 35.2 Causes of Ionized Hypocalcemia in the ICU


Fat embolism

Blood transfusions (15%)

Magnesium depletion (70%)

Cardiopulmonary bypass



Renal insufficiency (50%)

Aminoglycosides (40%)

Sepsis (30%)

Cimetidine (30%) Heparin (10%) Theophylline (30%) Numbers in parentheses show the frequency of ionized hypocalcemia reported in each condition.


Sepsis Sepsis is a common cause of hypocalcemia in the ICU (8,9). The mechanism is unclear, but it may involve an increase in calcium binding to albumin caused by elevated levels of circulating free fatty acids. Hypocalcemia is independent of the vasodilation that accompanies sepsis (9), and thus the clinical significance of the hypocalcemia in sepsis is unclear.

Alkalosis As mentioned earlier, alkalosis promotes the binding of calcium to albumin and can reduce the fraction of ionized calcium in blood. Symptomatic hypocalcemia is more common with respiratory alkalosis than with metabolic alkalosis. Infusions of sodium bicarbonate can also be accompanied by ionized hypocalcemia because calcium directly binds to the infused bicarbonate.

Blood Transfusions Ionized hypocalcemia has been reported in 20% of patients receiving blood transfusions (8). The mechanism is calcium binding by the citrate anticoagulant in banked blood. Hypocalcemia from blood transfusions usually is transient, and resolves when the infused citrate is metabolized by the liver and kidneys (8). In patients with renal or hepatic failure, a more prolonged hypocalcemia can result. Although hypocalcemia from blood transfusions could impede blood coagulation, this is not considered to be a significant effect and calcium infusions are no longer recommended in massive blood transfusions.

Drugs A number of drugs can bind calcium and promote ionized hypocalcemia ( 8). The ones most often used in the ICU are aminoglycosides, cimetidine, heparin, and theophylline.

Renal Failure Ionized hypocalcemia can accompany renal failure as a result of phosphate retention and impaired conversion of vitamin D to its active form in the kidneys. The treatment is aimed at lowering the phosphate levels in blood with antacids that block phosphorus absorption in the small bowel. However, the value of this practice is unproven. The acidosis in renal failure can decrease the binding of calcium to albumin, so hypocalcemia in renal failure does not imply ionized hypocalcemia.

Pancreatitis Severe pancreatitis can produce ionized hypocalcemia through several mechanisms. The prognosis is adversely affected by the appearance of hypocalcemia (10), although a causal relationship has not been proven. P.643

Clinical Manifestations The clinical manifestations of hypocalcemia are related to enhanced cardiac and neuromuscular excitability and reduced contractile force in cardiac muscle and vascular smooth muscle.

Neuromuscular Excitability Hypocalcemia can be accompanied by tetany (of peripheral or laryngeal muscles), hyperreflexia, paresthesias, and seizures (11). Chvostek's and Trousseau's signs are often listed as manifestations of hypocalcemia. However, Chvostek's sign is nonspecific (it is present in 25% of normal adults), and Trousseau's sign is insensitive (it can be absent in 30% of patients with hypocalcemia) (12).

Cardiovascular Effects The cardiovascular complications of hypocalcemia include hypotension, decreased cardiac output, and ventricular ectopic activity. These complications are rarely seen in mild cases of ionized hypocalcemia (i.e., ionized calcium 0.8 to 1.0 mmol/L). However, advanced stages of ionized hypocalcemia (i.e., ionized calcium less than 0.65 mmol/L)

can be associated with heart block, ventricular tachycardia and refractory hypotension (8).

Calcium Replacement Therapy The treatment of ionized hypocalcemia should be directed at the underlying cause of the problem. However, symptomatic hypocalcemia is considered a medical emergency (8), and the treatment of choice is intravenous calcium. The calcium solutions and dosage recommendations for intravenous calcium replacement are shown in Table 35.3. TABLE 35.3 Intravenous Calcium Replacement Therapy


Elemental Calcium

Unit Volume


10% Calcium chloride

27 mg (1.36 mEq)/mL

10-mL ampules

2000 mOsm/L

10% Calcium gluconate

9 mg (0.46 mEq)/mL

10-mL ampules

680 mOsm/L

For symptomatic hypocalcemia: 1. Infuse calcium into a large central vein if possible. If a peripheral vein is used, calcium gluconate should be used. 2. Give a bolus dose of 200 mg elemental calcium (8 mL of 10% calcium chloride or 22 mL of 10% calcium gluconate) in 100 mL isotonic saline over 10 minutes. 3. Follow with a continuous infusion of 1–2 mg elemental calcium per kg per hour for 6–12 hours.


Calcium Salt Solutions The two most popular calcium solutions for intravenous use are 10% calcium chloride and 10% calcium gluconate. Both solutions have the same concentration of calcium salt (i.e., 100 mg/mL), but calcium chloride contains three times more elemental calcium than calcium gluconate. One 10-mL ampule of 10% calcium chloride contains 272 mg (13.6 mEq) of elemental calcium, whereas one 10-mL ampule of 10% calcium gluconate contains only 90 mg (4.6 mEq) of elemental calcium (8).

Dosage Recommendations

The intravenous calcium solutions are hyperosmolar and should be given through a large central vein if possible. If a peripheral vein is used, calcium gluconate is the preferred solution because of its lower osmolarity (Table 35.3). A bolus dose of 100 mg elemental calcium (diluted in 100 mL isotonic saline and given over 5–10 minutes) should raise the total serum calcium by 0.5 mg/dL, but levels will begin to fall after 30 minutes (8). Therefore, the bolus dose of calcium should be followed by a continuous infusion at a dose rate of 0.5 to 2 mg/kg/hr (elemental calcium) for at least 6 hours. Individual responses will vary, so calcium dosing should be guided by the level of ionized calcium in blood (8).

Caution Intravenous calcium replacement can be risky in select patient populations. Calcium infusions can promote vasoconstriction and ischemia in any of the vital organs (13). The risk of calcium-induced ischemia should be particularly high in patients with low cardiac output who are already vasoconstricted. In addition, aggressive calcium replacement can promote intracellular calcium overload, which can produce a lethal cell injury ( 14), particularly in patients with circulatory shock. Because of these risks, calcium infusions should be used judiciously. Intravenous calcium is indicated only for patients with symptomatic hypocalcemia or an ionized calcium level below 0.65 mmol/L (8).

Maintenance Therapy The daily maintenance dose of calcium is 2–4 g in adults. This can be administered orally using calcium carbonate (e.g., Oscal) or calcium gluconate tablets (500 mg calcium per tablet).

Hypercalcemia Hypercalcemia is not nearly as common as hypocalcemia: it is reported in less than 1% of hospitalized patients (15). In 90% of cases, the underlying cause is hyperparathyroidism or malignancy (16,17). Less common causes include prolonged immobilization, thyrotoxicosis, and drugs (lithium, thiazide diuretics). Malignancy is the most common cause of severe hypercalcemia (i.e., total serum calcium above 14 mg/dL or ionized calcium above 3.5 mmol/L) (17). P.645

Clinical Manifestations The manifestations of hypercalcemia usually are nonspecific and can be categorized as follows (16): 1. 2. 3. 4.

Gastrointestinal (GI): nausea, vomiting, constipation, ileus, and pancreatitis Cardiovascular: hypovolemia, hypotension, and shortened QT interval Renal: polyuria and nephrocalcinosis Neurologic: confusion and depressed consciousness, including coma

These manifestations can become evident when the total serum calcium rises above 12 mg/dL (or the ionized calcium rises above 3.0 mmol/L), and they are almost always present when the serum calcium is greater than 14 mg/dL (or the ionized calcium is

above 3.5 mmol/L) (17).

Management Treatment is indicated when the hypercalcemia is associated with adverse effects, or when the serum calcium is greater than 14 mg/dL (ionized calcium above 3.5 mmol/L). The management of hypercalcemia is summarized in Table 35.4 (1,16,17).

Saline Infusion Hypercalcemia usually is accompanied by hypercalciuria, which produces an osmotic diuresis. This eventually leads to hypovolemia, which reduces calcium excretion in the urine and precipitates a rapid rise in the serum calcium. Therefore, volume infusion to correct hypovolemia and promote renal calcium excretion is the first goal of management for hypercalcemia. Isotonic saline is recommended for the volume infusion because natriuresis promotes renal calcium excretion.

Furosemide Saline infusion will not return the calcium to normal levels. This requires the addition of furosemide (40 to 80 mg IV every 2 hours) to further promote urinary calcium excretion. The goal is an hourly urine output of 100 to 200 mL/minute. The hourly urine output must be replaced with isotonic saline. Failure to replace urinary volume losses is counterproductive, and favors a return to hypovolemia.

Calcitonin Although saline and furosemide will correct the hypercalcemia acutely, this approach does not treat the underlying cause of the problem, which (in malignancy) is enhanced bone resorption. Calcitonin is a naturally occurring hormone that inhibits bone resorption. It is available as salmon calcitonin, which is given subcutaneously or intramuscularly in a dose P.646 of 4 U/kg every 12 hours. The response is rapid (onset within a few hours), but the effect is mild (the maximum drop in serum calcium is 0.5 mmol/L). TABLE 35.4 Management of Severe Hypercalcemia




Isotonic saline


Initial treatment of choice. Goal is rapid correction of hypovolemia.


40–80 mg IV every 2 hours

Add to isotonic saline to maintain a urine output of 100–200 mL/hr.


4 Units/kg IM or SC every 12 hours

Response is evident within a few hours. Maximum drop in serum calcium is only 0.5 mmol/L.


200 mg IV daily in 2–3 divided doses

Used as an adjunct to calcitonin.


More potent than calcitonin, but complete response requires 4–10 days.


90 mg IV over 2 hours

Reduce dose to 60 mg in renal impairment.


4 mg IV over 15 minutes

Equivalent to pamidronate in efficacy.

25 µg/kg IV over 4 hours; can repeat every 2 hours

More rapid effect than pamidronate, but potential for toxic side effects limits the use of this agent.


Hydrocortisone Corticosteroids can reduce the serum calcium by impeding the growth of lymphoid neoplastic tissue and enhancing the actions of vitamin D. Steroids are usually combined with calcitonin and can be particularly useful in the hypercalcemia associated with multiple myeloma or renal failure (1,16,17). The standard regimen uses hydrocortisone, 200 mg IV daily in 2 or 3 divided doses.

Intravenous Bisphosphonates Calcitonin can be used for rapid reduction of serum calcium, but the mild response will not keep the calcium in the normal range. A group of compounds known as bisphosphonates (pyrophosphate derivatives) are more potent inhibitors of bone resorption and maintain a normal serum calcium. However, their onset of action is delayed, and thus they are not useful when rapid control of serum calcium is desired. P.647 Zoledronate (4 mg over 15 minutes) or Pamidronate (90 mg over 2 hours) are the bisphosphonates of choice for the management of severe hypercalcemia ( 17). The peak effect is seen in 2 to 4 days, and serum calcium normalizes within 4–7 days in 60–90% of cases. The dose may be repeated in 4–10 days, if necessary.

Plicamycin Plicamycin (formerly mithramycin) is an antineoplastic agent that inhibits bone resorption. It is similar to the bisphosphonates in that it is more potent than calcitonin but has a delayed onset of action. The dose is 25 mg/kg (intravenously over 4–6 hours), which can be repeated in 24–48 hours if necessary (17). Because of the potential for serious side effects (e.g., bone marrow suppression), plicamycin has largely been replaced by pamidronate.

Dialysis Dialysis (hemodialysis or peritoneal dialysis) is effective in removing calcium in patients with renal failure (17).

Phosphorus The average adult has 500–800 g of phosphorus (18,19). Most is contained in organic molecules such as phospholipids and phosphoproteins, and 85% is located in the bony skeleton. The remaining 15% in soft tissues is present as free, inorganic phosphorus. Unlike calcium, inorganic phosphorus is predominantly intracellular in location, where it participates in glycolysis and high energy phosphate production. The normal concentration of inorganic phosphorus in plasma is shown in Table 35.1.

Hypophosphatemia Hypophosphatemia (serum PO 4 less than 2.5 mg/dL or 0.8 mmol/L) is reported in 17 to 28% of critically ill patients (20,21) and can be the result of an intracellular shift of phosphorus, an increase in the renal excretion of phosphorus, or a decrease in phosphorus absorption from the GI tract. Most cases of hypophosphatemia are due to movement of PO4 into cells.

Predisposing Conditions

Glucose Loading The movement of glucose into cells is accompanied by a similar movement of PO 4 into cells, and if the extracellular content of PO 4 is marginal, this intracellular PO 4 shift can result in hypophosphatemia. Glucose loading is the most common cause of hypophosphatemia in hospitalized patients (20,22), usually seen during refeeding in alcoholic, malnourished, or debilitated patients. It can occur with oral feedings, enteral tube feedings, or with total parenteral nutrition. The influence of parenteral P.648 nutrition on serum PO4 levels is shown in Figure 35.2. Note the gradual decline in the serum PO4 and the severe degree of hypophosphatemia (serum PO 4 90%) but a very low specificity (15 to 40%) for the diagnosis of pneumonia (15). This

means that a negative culture of a tracheal aspirate can be used to exclude the diagnosis of pneumonia, but a positive culture cannot be used to confirm the presence of pneumonia. The poor predictive value of positive cultures is due to contamination of tracheal aspirates with secretions from the mouth and upper airways. The diagnostic accuracy of tracheal aspirates can be improved by screening the specimens with microscopic visualization to include only specimens originating from the lower airways, and then performing quantitative cultures on the screened specimens. This is explained in the following sections.

Microscopic Analysis The cells identified in Figure 41.3 can help to determine if aspirated secretions originate in the upper or lower airways, and also if there is evidence of infection. Each type of cell can be identified and interpreted as follows. The squamous epithelial cells that line the oral cavity are large and flattened with abundant cytoplasm and a small nucleus (see Fig. 41.3). The presence of more than 25 squamous epithelial cells per low-power field (× 100) indicates that the specimen is contaminated with mouth secretions (16). If there is evidence of contamination, the specimen should be discarded. Lung macrophages are large, oval-shaped cells with a granular cytoplasm and a small, eccentric nucleus (Fig. 41.3). The size of the nucleus P.755 in a macrophage is roughly the same size as a neutrophil. Although macrophages can inhabit the airways (17), the predominant home of the macrophage is the distal airspaces. Therefore, the presence of macrophages, regardless of the number, indicates that the specimen is from the lower respiratory tract. Figure 41.3 Microscopic appearance (magnification ×400) of bronchial brushings from a ventilator-dependent patient. The paucity of squamous epithelial cells and the presence of alveolar macrophages both confirm that the specimen is from the distal airways.

View Figure

The presence of neutrophils in respiratory secretions is not evidence of infection because neutrophils can make up 20% of the cells recovered from a routine mouthwash (17). The neutrophils should be present in abundance to indicate infection. More than 25 neutrophils per low-power field (× 100) can be used as evidence of infection (18). When an infection is evident, a search for macrophages and squamous epithelial cells will help to determine if the infection is in the upper airways (tracheobronchitis) or the lower airways (pneumonia). When a tracheal aspirate shows evidence of lower airways infection (i.e., abundant neutrophils and macrophages with few epithelial cells), the specimen is suitable for culture.

Quantitative Cultures Quantitative cultures of tracheal aspirates produce fewer false-positive results than qualitative cultures. To perform quantitative cultures, the tracheal aspirate should be collected in a sterile trap without adding saline or lidocaine (the latter can inhibit microbial growth). When a volume of at least 1 mL is collected, the specimen is sent to the laboratory P.756 (make sure you specify that you want quantitative cultures) where it is vortexed (agitated) in isotonic saline. A sample is then collected with a 0.01mL loop and the sample is placed on a culture plate for incubation. For quantitative cultures, growth on the plate is reported as follows: 10 colonies is reported as 10 3 colony-forming units per mL (CFU/mL), 100 colonies is 10 4 CFU/mL, 1,000 colonies is 105 CFU/mL, and more than 1,000 colonies is 106 CFU/mL (this technique might vary in different laboratories, but the principle is the same). TABLE 41.3 Quantitative Cultures for the Diagnosis of Pneumonia in Ventilator-Dependent Patients




105 to 106


104 to 105

Sensitivity (mean)




Specificity (mean)




Most Sensitive

Most Specific

Most Accurate

Diagnostic Threshold (cfu/mL)

Relative Performance

Abbreviations: TA = tracheal aspirates, PSB = protected specimen brushings, BAL = bronchoalveolar lavage. From References (1,15,21).

For quantitative cultures of tracheal aspirates, the threshold growth for the diagnosis of pneumonia is 10 5 to 106 CFU/ per mL (the lower threshold is sometimes used for patients who are on antibiotic therapy when the cultures are performed). Studies using these thresholds have shown a (mean) sensitivity and specificity of 76% and 75%, respectively, for the diagnosis of pneumonia (see Table 41.3) (1,15). Comparing these results to the sensitivity and specificity of qualitative cultures mentioned earlier (i.e., sensitivity >90% and specificity =40%) shows that quantitative cultures of tracheal aspirates are less sensitive but much more specific than qualitative cultures for the diagnosis of


Protected Specimen Brush Aspiration of secretions through a bronchoscope produces a high rate of false-positive cultures because the bronchoscope picks up contaminants as it is passes through the upper respiratory tract (19). To eliminate this problem, a specialized brush called a protected specimen brush (PSB) was developed to collect uncontaminated secretions from the distal airways. The brush sits in the inner lumen of a catheter-over-catheter device that has a gelatin plug at the distal end. When the device is advanced through the bronchoscope, the gelatin plug protects the brush from contamination with upper airways secretions. When the bronchoscope is advanced into the area of lung infiltration, the entire catheter device is advanced out of the bronchoscope and into the lower airways (see Fig. 41.4). The inner catheter is advanced until it knocks off the gelatin plug (which dissolves without harming the patient), and the brush is then advanced into the distal airways to collect the specimen. After vigorous brushing, the brush is retracted into the inner cannula, the inner cannula is retracted P.757 into the outer cannula, and the entire device is retracted through the bronchoscope. Figure 41.4 The protected specimen brush (PSB) technique for obtaining uncontaminated secretions from the lower airways.

View Figure

Quantitative Cultures Using sterile technique, the brush is severed from its wire and is placed in 1 mL of transport medium. In the microbiology lab, the brush is vortexed in the transport medium to disperse microorganisms. The specimen is P.758 then processed in the same fashion as described for tracheal aspirates. Growth of 10 3 CFU/mL is the threshold for the presence of infection (pneumonia) (1). The reported sensitivity and specificity of PSB cultures are shown in Table 41.3 (1). A positive culture

result has a relatively low sensitivity (66%) but a high specificity (90%) for the diagnosis of pneumonia. This means that a negative PSB culture does not exclude the presence of pneumonia, but a positive PSB culture confirms the presence of pneumonia with 90% certainty.

Bronchoalveolar Lavage Bronchoalveolar lavage (BAL) is performed by wedging the bronchoscope in a distal airway and performing a lavage with sterile isotonic saline. A minimum lavage volume of 120 mL is recommended for adequate sampling of the lavaged lung segment (19), and this is achieved by performing a series of 6 lavages using 20 mL for each lavage. The same syringe is used to introduce the fluid and aspirate the lavage specimen (only 25% or less of the volume instilled will be returned via aspiration).

Quantitative Cultures The first lavage is usually discarded, and the remainder of the lavage fluid is pooled and sent to the microbiology lab, where it is centrifuged and processed in the same manner as described for tracheal aspirates. The threshold for a positive BAL culture is 10 4 to 105 CFU/mL (the lower threshold is for patients who are on antibiotic therapy when the procedure is performed) (1). The reported sensitivity and specificity of BAL cultures are shown in Table 41.3 (1,20). Neither the sensitivity nor the specificity exceeds both of the diagnostic methods in Table 41.3, but when sensitivity and specificity are considered together, BAL cultures have the highest overall accuracy for the diagnosis of pneumonia.

Intracellular Organisms Inspection of BAL specimens for intracellular organisms can help in guiding initial antibiotic therapy until culture results are available. When intracellular organisms are present in more than 3% of the cells in the lavage fluid, the likelihood of pneumonia is over 90% (21). Unfortunately, this is not done on a routine Gram stain, but requires special processing and staining in the microbiology lab (see references 21 and 22 for the methodology).

BAL without Bronchoscopy The limited availability of bronchoscopy on a 24-hour basis has led to the introduction of a non-bronchoscopic technique for BAL where specialized catheters are inserted through a tracheal tube and advanced blindly until wedged in the lower airways. A variety of catheters and techniques have been used (for examples of two techniques, see references 23 and 24), and the method has proven safe and effective when performed by respiratory therapists (23). In general, cultures obtained by nonbronchoscopic P.759 and bronchoscopic BAL have shown equivalent sensitivities and specificities for the diagnosis of pneumonia (1,25), and there is evidence that nonbronchoscopic BAL is better tolerated than bronchoscopic BAL (24). However, there is concern about adopting nonbronchoscopic procedures to diagnose pneumonia because of the lack of standardization of the techniques in clinical studies (26).

Which Diagnostic Method is Best?

There is little agreement on which method should be preferred for the diagnosis of pneumonia. The following statements include some pertinent observations. 1. The diagnostic yield from all culture methods is adversely affected by ongoing antibiotic therapy (1). Therefore, when possible, cultures should be obtained before antibiotics are started. 2. Tracheal aspirates should be screened by microscopic examination, and the specimens should be discarded if there is evidence of contamination with secretions from the mouth and upper airways. 3. Quantitative cultures of tracheal aspirates are preferred to qualitative cultures because they have a higher specificity, and thus are more likely to identify a pneumonia and the responsible pathogen(s). 4. Treatment based on cultures of tracheal aspirates will result in excessive use of antibiotics because cultures of tracheal aspirates are most likely to produce false-positive results (1,27). 5. Most studies show that the mortality in ventilator-associated pneumonia is not influenced by the diagnostic method (1,27). In other words, there is no survival benefit with the more invasive methods (protected specimen brushings and bronchoalveolar lavage) compared to the relatively simple method of aspirating secretions through an endotracheal or tracheostomy tube. Based on the absence of a survival benefit with the invasive diagnostic methods, many prefer tracheal aspirates for the diagnostic approach to pneumonia, even though this practice will result in excessive use of antibiotics (1). Tracheal aspirates should, however, be screened by microscopic inspection and cultured using quantitative techniques. Qualitative cultures of tracheal aspirates are useful only when there is no growth, which can be used to exclude the presence of pneumonia.

Parapneumonic Effusions Pleural effusions are present in up to 50% of bacterial pneumonias (28), and these parapneumonic effusions should be evaluated in the following situations (1): when the pleural effusion is large, when the patient appears toxic, and when the patient is not improving on antibiotic therapy. If the effusion is loculated (i.e., it does not move with change in body position), computed tomography or bedside ultrasound can be P.760 used to mark the location and depth of the fluid. In addition to Gram stain and culture, the pleural fluid glucose concentration and pH should be measured. Classification of the fluid as a transudate or exudate (by pleural fluid protein and LDH levels) is unnecessary because this does not reliably identify infection. TABLE 41.4 Management of Parapneumonic Effusions

Clinical Finding

Immediate Drainage? Yes

Air–fluid level




Grossly purulent fluid



Pleural fluid pH: 7.2


Pleural fluid glucose: 40 mg/dL


Indications for Drainage The indications for immediate drainage of parapneumonic effusions are listed in Table 41.4. The radiographic criteria for drainage include the presence of an air–fluid level in the effusion, or a hydropneumothorax (both are signs of a bronchopleural fistula). Chemical criteria for drainage include a pleural fluid glucose concentration below 40 mg/dL (2.4 µmol/L) or a pleural fluid pH below 7.0 (28). If a patient with a parapneumonic effusion improves clinically on antibiotics, there is little need for further evaluation or drainage of the effusion.

Early Antimicrobial Therapy There is a definite tendency to begin antibiotics at the earliest hint of a pneumonia in the ICU. According to one study, antibiotic treatment for pneumonia accounts for half of all antibiotic use in the ICU, but 60% of antibiotic use for pneumonias involve suspected pneumonias that are not confirmed by bacteriologic studies (29). The rush to antibiotics in

cases of suspected pneumonia is fueled by studies showing that the mortality in ventilator-associated pneumonia is increased when there is a delay in starting appropriate antibiotic therapy (30). However, there are also studies showing that ventilator-associated pneumonia does not increase mortality (2,31,32), so the survival benefit of the rush to antibiotics is not firmly established. (See the last section of this chapter for a comment on this situation.) P.761

Suggested Strategy One early antibiotic strategy that is appealing is to administer antibiotics immediately (i.e., when pneumonia is first suspected) for patients who are immunocompromised, or have evidence of severe sepsis (i.e., multiorgan dysfunction) or septic shock. Otherwise, hold antibiotics until you collect specimens from the respiratory tract for Gram stain and culture. Then begin antibiotic therapy based on the appearance of the Gram stain, or using recommended empiric regimens (described next) but stop the antibiotics in 2 to 3 days if the cultures do not confirm the presence of pneumonia (33). There will be a tendency to continue antibiotics despite sterile cultures in patients with progressive respiratory insufficiency, but these patients are likely to have ARDS and not pneumonia. If respiratory tract cultures are obtained when patients are not receiving antibiotics, continuing antibiotics in the face of sterile cultures is rarely justified.

Empiric Antibiotic Therapy The choice of empiric antibiotics should be dictated by the likelihood that the patient is colonized with Staphylococcus aureus and gram-negative enteric pathogens (the pathogens listed in Table 41.1). The characteristics of patients who are likely and unlikely to be colonized with these pathogens are shown in Table 41.5, along with the recommended empiric antibiotic regimens for each type of patient. The recommended starting doses for each antibiotic are shown in Table 41.6.

Colonization Unlikely The typical patients who are not colonized by S. aureus and gram-negative pathogens have been admitted recently (within 5 days) from home, have no debilitating chronic illness (including renal failure that requires dialysis), and have no other hospital admissions in the past 3 months (1). Patients like this can be treated with a single antibiotic like ceftriaxone or a fluoroquinolone (levofloxacin or moxifloxacin), as shown in Table 41.5. This treatment is primarily directed at pneumococci, including penicillin-resistant strains, and is similar to the treatment of community-acquired pneumonia (34).

Colonization Likely Colonization with S. aureus and gram-negative enteric pathogens is likely when a patient has been in the hospital for 5 days or longer, or when the patient is a nursing home resident, has a chronic debilitating illness, or has had other hospital admissions within the past few months. The empiric antibiotic regimen for these patients, which is shown in Table 41.5, is designed to cover staphylococci and gram-negative enteric pathogens, and is particularly designed for Pseudomonas aeruginosa and methicillin-resistant S. aureus (MRSA). There is no coverage for fungi, and coverage for anaerobes is not a priority because anaerobes are not considered to be important pathogens in

ventilator-associated pneumonia (35). Specific choices of antibiotics will be guided by the profile of nosocomial pathogens and resistance patterns in individual ICUs. TABLE 41.5 Empiric Treatment for Ventilator-Associated Pneumonia Based on Likelihood of Colonization with Pathogens in TABLE 41.1 †

Colonization Unlikely Type of Patient:

Colonization Likely Type of Patient:

Admitted less than 5 days ago and

Admitted more than 5 days ago, or

Admitted from home, and

Admitted from a nursing home, or

No other admissions in past 3 months, and

Other admissions in the past 3 months, or

Not a dialysis patient.

A dialysis patient.

Empiric Antibiotics:

Empiric Antibiotics:

Ceftriaxone, or

Pipericillin/tazobactam, or

A fluoroquinolone (levofloxacin, moxifloxacin, or ciprofloxacin)

Imipenem or meropenem, or Ceftazidime or cefepime plus Ciprofloxacin or levofloxacin*, or An aminoglycoside* plus Vancomycin or linezolid‡

Vancomycin or linezolid‡ *Benefit is questionable (see text). ‡When colonization with methicillin-resistant S. aureus is known or likely. †

From the American Thoracic Society and Infectious Disease Society of America: Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005; 171:388–416.


Double Coverage for Pseudomonas? The empiric regimen for patients likely to be colonized with gram-negative pathogens includes a second antibiotic for gram-negative coverage (a fluoroquinolone or aminoglycoside), which is designed to provide double coverage for pneumonias caused by Pseudomonas aeruginosa. This recommendation is questionable and is based on one study that showed improved survival in patients with Pseudomonas bacteremia (many of whom did not have pneumonia) when double antibiotic coverage was used instead of monotherapy (36). However, another study shows no difference in survival when double coverage for Pseudomonas is compared to single drug therapy in patients with serious gram-negative infections (37). Furthermore, empiric double coverage for Pseudomonas pneumonia is not justified based on the following reasoning: (1) pneumonia is expected in only about 30% of ICU patients with suspected pneumonia, (2) in the patients with pneumonia, Pseudomonas is expected in only 20% of cases, and (3) in the patients with Pseudomonas pneumonia, bacteremia is expected in only about 25% of cases. If these estimates are combined, then Pseudomonas bacteremia (the only condition with evidence of benefit from double coverage) is expected in only 1.5% of ICU patients with suspected pneumonia. TABLE 41.6 Recommended Starting Doses for Empiric Antibiotics †


Intravenous Dosage*

ß-Lactam/ß-lactamase inhibitor Pipericillin/tazobactam

4.5 grams every 6 hours

Carbepenems Imipenem

1 gram every 8 hours


1 gram every 8 hours

Antipseudomonal cephalosporins Cefepime

1–2 grams every 8–12 hours


2 grams every 8 hours

Antipseudomonal quinolones Levofloxacin

750 mg once daily


400 mg every 8 hours

Aminoglycosides Gentamicin

7 mg/kg daily


7 mg/kg daily


20 mg/kg daily

Antistaphylococcal agents Vancomycin

15 mg/kg every 12 hours


600 mg/kg every 12 hours


600 mg/kg every 12 hours

*For patients with normal renal and hepatic function. †From

the American Thoracic Society and Infectious Disease Society of America: Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005; 171:388–416.


Recommendation For patients likely to be colonized with S. aureus and gram-negative pathogens, empiric antibiotic therapy for most patients will be adequate using imipenem or imipenem plus vancomycin. Imipenem has a very broad spectrum of activity, and is active against staphylococci (methicillin-sensitive strains), and gram-negative enteric pathogens, including Pseudomonas aeruginosa. The addition of vancomycin is necessary only if methicillin-resistant Staphylococcus aureus is a concern. Linezolid is gaining popularity as an alternative to vancomycin because of increasing vancomycin failures for treating MRSA pneumonias (1). The second antipseudomonal antibiotic should not be necessary unless the Gram stain of respiratory secretions shows a preponderance of gram-negative bacilli and the patient is immunocompromised (see Chapter 43 for the management of immunocompromised patients in the ICU).

Duration of Antibiotic Therapy Empiric antibiotics regimens will be adjusted according to the results of quantitative cultures. For pneumonias documented by culture, the P.764 traditional duration of antibiotic therapy has been 14 to 21 days (1). However, one large study (including 50 ICUs) has shown that 8 days of antibiotic therapy for ventilator-associated pneumonia is associated with the same mortality and risk of recurrent infection as 15 days of therapy (38), and the popular opinion at present is that one week of antibiotic therapy is adequate for most patients with ventilator-associated pneumonia.

Antibiotic Prophylaxis Efforts to prevent nosocomial pneumonia continue to be neglected despite evidence that some measures are effective. One of the effective measures involves the topical application of an antimicrobial paste to the oral mucosa to prevent colonization of the oropharynx with pathogenic organisms (39). The preparation most often used is a methylcellulose paste (Orabase, Squibb Pharmaceuticals) containing 2% polymyxin, 2% tobramycin, and 2% amphotericin B, which is applied to the inside of the mouth with a

gloved finger every 6 hours (39). As shown in Figure 4.6 (Chapter 4), this regimen of oral decontamination decreases the incidence of ventilator-associated pneumonia by about 2/3 (39). Results like this should not be ignored in the approach to pneumonia in the ICU.

A Final Word Reports that mortality is not increased by ventilator-associated pneumonia (2,31,32) deserve much more attention in discussions of how to manage pneumonia in the ICU. For example, if mortality is not increased by ventilator-associated pneumonias, then mortality rate should not be used as an end-point for evaluating diagnostic or therapeutic approaches to these pneumonias (as it often is). It also means that ICU-acquired pneumonia is overhyped as a life-threatening condition that requires immediate and aggressive therapy. Because there is no gold-standard method for identifying ICU-acquired pneumonia other than post-mortem examination (1), it is possible that studies showing a lack of impact on survival included many false-positive diagnoses of pneumonia. However, this observation is still relevant because these studies used methods that we all use to diagnose pneumonia and to justify antibiotic therapy. If what we think is pneumonia has no impact on survival, then the problems mentioned earlier still apply. For now, this is just another problem in a long list of problems associated with pneumonia in the ICU.

References Clinical Practice Guidelines 1. American Thoracic Society and Infectious Disease Society of America. Guidelines for the management of adults with hospital-acquired, ventilator-associated, and healthcare-associated pneumonia. Am J Respir Crit Care Med 2005;171:388–416. Full TextBibliographic Links P.765

General Features 2. Bregeon F, Cias V, Carret V, et al. Is ventilator-associated pneumonia an independent risk factor for death? Anesthesiology 2001;94:554–560.

3. Estes RJ, Meduri GU. The pathogenesis of ventilator-associated pneumonia, I: mechanisms of bacterial transcolonization and airway inoculation. Intensive Care Med 1995;21:365–383. Bibliographic Links 4. Wunderink RG. Clinical criteria in the diagnosis of ventilator-associated pneumonia. Chest 2000;117:191S–194S.

Full TextBibliographic Links 5. Fagon JY, Chastre J, Hance AJ, et al. Detection of nosocomial lung infection in ventilated patients: use of a protected specimen brush and quantitative culture techniques in 147 patients. Am Rev Respir Dis 1988;138:110–116. Bibliographic Links 6. Timsit JF, Misset B, Goldstein FW, et al. Reappraisal of distal diagnostic testing in the diagnosis of ICU-acquired pneumonia. Chest 1995;108:1632–1639. Bibliographic Links 7. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128–140. Full TextBibliographic Links 8. Nsieir S, Di Pompeo C, Pronnier P, et al. Nosocomial tracheobronchitis in mechanically ventilated patients: incidence, aetiology, and outcome. Eur Respir J 2002;20:1483–1489. Bibliographic Links 9. Pistolesi M, Miniati M, Milne ENC, et al. Measurement of extravascular lung water. Intensive Care World 1991;8:16–21. Bibliographic Links 10. Caldwell A, Glauser FL, Smith WP, et al. The effects of dehydration on the radiologic and pathologic appearance of experimental canine segmental pneumonia. Am Rev Respir Dis 1975;112:651–659. Bibliographic Links 11. Louthan FB, Meduri GU. Differential diagnosis of fever and pulmonary densities in mechanically ventilated patients. Semin Resp Infect 1996;11:77–95.

12. Singh N, Falestiny MN, Rogers P, et al. Pulmonary infiltrates in the surgical ICU. Chest 1998;114:1129–1136. Bibliographic Links 13. Bauer TT, Torres A. Acute respiratory distress syndrome and nosocomial pneumonia. Thorax 1999;54:1036–1040. Bibliographic Links

Diagnostic Evaluation 14. Luna CM, Videla A, Mattera J, et al. Blood cultures have limited value in predicting severity of illness and as a diagnostic tool in ventilator-associated

pneumonia. Chest 1999;116:1075–1084. Bibliographic Links 15. Cook D, Mandell L. Endotracheal aspiration in the diagnosis of ventilator-associated pneumonia. Chest 2000;117:195S–197S. Full TextBibliographic Links 16. Washington JA. Techniques for noninvasive diagnosis of respiratory tract infections. J Crit Illness 1996;11:55–62.

17. Rankin JA, Marcy T, Rochester CL, et al. Human airway macrophages. Am Rev Respir Dis 1992;145:928–933. Bibliographic Links 18. Wong LK, Barry AL, Horgan S. Comparison of six different criteria for judging the acceptability of sputum specimens. J Clin Microbiol 1982;16:627–631. Bibliographic Links 19. Meduri GU, Chastre J. The standardization of bronchoscopic techniques for ventilator-associated pneumonia. Chest 1992;102:557S–564S. Bibliographic Links 20. Torres A, El-Ebiary M. Bronchoscopic BAL in the diagnosis of ventilator-associated pneumonia. Chest 2000;117:198S–202S. Full TextBibliographic Links P.766 21. Veber B, Souweine B, Gachot B, et al. Comparison of direct examination of three types of bronchoscopy specimens used to diagnose nosocomial pneumonia. Crit Care Med 2000;28:962–968. Ovid Full TextBibliographic Links 22. Chastre J, Fagon JY, Domart Y, et al. Diagnosis of nosocomial pneumonia in intensive care unit patients. Eur J Clin Microbiol Infect Dis 1989;8:35–39. Bibliographic Links 23. Kollef MH, Bock KR, Richards RD, et al. The safety and diagnostic accuracy of minibronchoalveolar lavage in patients with suspected ventilator-associated pneumonia. Ann Intern Med 1995;122:743–748. Ovid Full TextFull TextBibliographic Links 24. Perkins GD, Chatterjee S, Giles S, et al. Safety and tolerability of nonbronchoscopic lavage in ARDS. Chest 2005;127:1358–1363. Full TextBibliographic Links

25. Campbell CD Jr. Blinded invasive diagnostic procedures in ventilator-associated pneumonia. Chest 2000;117:207S–211S. Full TextBibliographic Links 26. Fujitani S, Yu V. Diagnosis of ventilator-associated pneumonia: focus on nonbronchoscopic techniques (nonbronchoscopic lavage, including mini-BAL, blinded protected specimen brush, and blinded bronchial sampling) and endotracheal aspirates. J Intensive Care Med 2006;21:17–21. Bibliographic Links 27. Shorr AF, Sherner JH, Jackson WL, et al. Invasive approaches to the diagnosis of ventilator-associated pneumonia: a meta-analysis. Crit Care Med 2005; 33:46–53. Ovid Full TextBibliographic Links 28. Light RW, Meyer RD, Sahn SA, et al. Parapneumonic effusions and empyema. Clin Chest Med 1985;6:55–62. Bibliographic Links

Early Antimicrobial Therapy 29. Bergmanns DCJJ, Bonten MJM, Gaillard CA, et al. Indications for antibiotic use in ICU patients: a one-year prospective surveillance. J Antimicrob Chemother 1997;111:676–685.

30. Iregui M, Ward S, Sherman G, et al. Clinical importance of delays in the initiation of appropriate antibiotic treatment for ventilator-associated pneumonia. Chest 2002;122:262–268. Full TextBibliographic Links 31. Rello J, Quintana E, Ausina A, et al. Incidence, etiology, and outcome of nosocomial pneumonia in mechanically ventilated patients. Chest 1991;100:439–444. Bibliographic Links 32. Papazian L, Bregeon F, Thirion X, et al. Effect of ventilator-associated pneumonia on mortality and morbidity. Am J Respir Crit Care 1996;154:91–97. Bibliographic Links 33. Singh N, Rogers P, Atwood CW, et al. Short-course empiric antibiotic therapy for patients with pulmonary infiltrates in the intensive care unit. Am J Respir Crit Care Med 2000;162:505–511. Bibliographic Links 34. American Thoracic Society. Guidelines for the management of adults with

community-acquired pneumonia. Am J Respir Crit Care Med 2001;163:1730–1754. Bibliographic Links 35. Marik P, Careau P. The role of anaerobes in patients with ventilator-associated pneumonia and aspiration pneumonia: a prospective study. Chest 1999;115:178–183. Bibliographic Links 36. Hilf M, Yu VL, Sharp J, et al. Antibiotic therapy for Pseudomonas aeruginosa bacteremia: outcome correlations in a prospective study of 200 patients. Am J Med 1989;87:540–546. Bibliographic Links 37. Cometta A, Baumgartner JD, Lew D, et al. Prospective, randomized comparison of imipenem monotherapy with imipenem plus netilmicin for treatment of severe infections in nonneutropenic patients. Antimicrob Agents Chemother 1994;38:1309–1313. Bibliographic Links P.767 38. Chastre J, Wolff M, Fagon J-Y, et al. Comparison of 8 vs 15 days of antibiotic therapy for ventilator-associated pneumonia in adults. JAMA 2003;290:2588–2598. Ovid Full TextBibliographic Links 39. Bergmans C, Bonten M, Gaillard C, et al. Prevention of ventilator-associated pneumonia by oral decontamination. Am J Respir Crit Care Med 2001;164:382–388. Bibliographic Links

Chapter 42 Sepsis from the Abdomen and Pelvis One of the recurring themes in this book is the importance of the gastrointestinal (GI) tract as a source of infection in critically ill patients. This chapter focuses on that theme, and describes the infectious risks at both ends of the GI tract, including the neighboring biliary tree. The last section of this chapter describes nosocomial infections in the urinary tract, with emphasis on infections associated with indwelling urethral catheters.

Acalculous Cholecystitis Acalculous cholecystitis is a condition that could be described as an ileus of the gallbladder. Although an uncommon condition, it can be fatal if not recognized and treated promptly (1).

Pathogenesis There are a number of conditions that predispose to acalculous cholecystitis. Most cases occur in association with multiple trauma and abdominal (nonbiliary) surgery. A number of mechanisms may be involved in the pathogenesis of acalculous cholecystitis, including ischemia (e.g., multiple trauma, shock), stasis (e.g., parenteral nutrition), and reflux of pancreatic secretions (e.g., opioid analgesics). In patients with immunodeficiency virus (HIV) infection, opportunistic pathogens like cytomegalovirus are often found on histologic examination of the gallbladder, but it is unclear if these organisms are the cause or the consequence of the cholecystitis (3).

Clinical Features The clinical manifestations of acalculous cholecystitis include fever, nausea and vomiting, abdominal pain, and right upper quadrant tenderness P.770 (Table 42.1). Abdominal findings can be minimal or absent, and fever may be the only presenting manifestation. Elevations in serum bilirubin, alkaline phosphatase, and amylase can occur but are variable (1,2). TABLE 42.1 Routine Clinical Evaluation in 143 Patients with Intra-Abdominal Abscesses

Clinical Finding

Frequency (%)

Physical Examination: Localized abdominal tenderness


Palpable abdominal mass


Chest Films: Pleural effusion


Basilar atelectasis


Abdominal Films: Extraluminal air or air-fluid level


Mechanical bowel obstruction


From Fry D. Noninvasive imaging tests in the diagnosis and treatment of intra-abdominal abscesses in the postoperative patient. Surg Clin North Am 1994; 74:693–709.

Diagnosis An ultrasound of the right upper quadrant often provides diagnostic information. Gallbladder sludge and distention of the gallbladder are common findings but can be nonspecific. More specific findings include a gallbladder wall thickness of at least 3.5 mm and submucosal edema (1,2). If ultrasound visualization is hampered, computed tomography (CT) scanning can provide useful information (1).

Management Prompt intervention is necessary to prevent progressive distention and rupture of the gallbladder. The latter complication has been reported in 40% of cases when diagnosis and treatment is delayed for 48 hours or longer after the onset of symptoms (1). The treatment of choice is cholecystectomy. In patients who are too moribund for surgery, percutaneous cholecystostomy is a suitable alternative. Empiric antibiotics are

often recommended, but the value of this practice is unproven.

Colonization of the GI Tract The GI tract can become a source of sepsis when overgrowth of pathogenic organisms occurs as a result of a change in the normal environment of the bowel lumen. This occurs in the upper GI tract when patients are given acid-suppressing drugs, and occurs in the lower GI tract when patients are given antibiotics. The following is a look at this colonization as a source of sepsis. Figure 42.1 Correlation between the organisms that most often colonize the upper GI tract and the organisms most often isolated in nosocomial infections in critically ill patients. (Data from Marshall JC, Christou NV, Meakins JL, et al. The gastrointestinal tract: the “undrained abscess” of multiple organ failure. Ann Surg 1993;218:111–119.Bibliographic Links)

View Figure


Gastric Colonization Gastric acid suppression and subsequent colonization of the upper GI tract is discussed in Chapter 4 (see Fig. 4.1 for a demonstration of the bactericidal actions of low pH). The pathogens that commonly colonize the stomach are the same pathogens that are commonly involved in nosocomial infections (4). This correlation is shown in Figure 42.1. Although it does not prove a causal relationship between gastric colonization and nosocomial infections, it does show that the upper GI tract serves as a reservoir for pathogens that are commonly involved in nosocomial sepsis.

Preventive Measures There are two practices that will reduce colonization of the upper GI tract, and both have proven to reduce the incidence of nosocomial infections. The first practice is to avoid the use of drugs that suppress gastric P.772 acidity (i.e., histamine H2 antagonists and proton pump inhibitors). Prophylaxis for stress ulcer bleeding is not necessary if the patient is eating or receiving enteral tube feedings. If

prophylaxis is desired, consider using sucralfate, a cytoprotective agent that does not increase the pH of gastric secretions. The relative advantages and disadvantages of using sucralfate instead of acid-inhibiting drugs are discussed in Chapter 4, and shown in Figure 4.4. The second practice that impedes colonization is the use of nonabsorbable antibiotics placed in the mouth and stomach. This practice is described in Chapter 4, and the results are shown in Figures 4.6 and 4.7. Despite the obvious benefits for reducing nosocomial infections, this practice is not popular in the United States (see Chapter 4 for more on this topic). However, the success of decontamination practices in reducing nosocomial infections is evidence that the GI tract is indeed an important source of nosocomial sepsis.

Clostridium Difficile Colitis Colonization with pathogenic organisms can also occur in the lower regions of the GI tract. The most troublesome intruder is Clostridium difficile, a spore-forming gram-positive anaerobic bacillus that is not a prominent bowel inhabitant in healthy subjects, but proliferates when the normal microflora of the lower GI tract is altered by antibiotic therapy (5,6). Clostridium difficile is not an invasive organism, but it elaborates cytotoxins that incite inflammation in the bowel mucosa. Severe cases of mucosal inflammation are accompanied by raised plaque-like lesions on the mucosal surface called pseudomembranes. These lesions are responsible for the term pseudomembranous colitis, which is used to describe advanced cases of C. difficile enterocolitis.

Epidemiology Although C. difficile is found in fewer than 5% of healthy adults in the community, it can be seen in as many as 40% of hospitalized patients (7). More than half of the patients who harbor C. difficile in their stool are asymptomatic (8). The organism is found primarily in patients receiving ongoing or recent (within 2 weeks) antibiotic therapy and in patients who are in close proximity to other patients who harbor the organism. C. difficile is readily transmitted from patient to patient by contact with contaminated objects (e.g., toilet facilities) and by the hands of hospital personnel (8). Strict adherence to the use of disposable gloves can significantly reduce the nosocomial transmission of C. difficile (9).

Clinical Manifestations The most common manifestations of symptomatic C. difficile infection are fever, abdominal pain, and watery diarrhea. Bloody diarrhea is seen in 5% to 10% of cases. Rarely, the enterocolitis can progress to toxic megacolon, which presents with abdominal distention, ileus, and clinical shock. This latter complication can be fatal and requires emergent subtotal colectomy (10). P.773

Laboratory Tests The diagnosis of C. difficile enterocolitis requires laboratory tests for the presence of the appropriate toxins in stool. Stool cultures for C. difficile are unreliable because they do not distinguish toxigenic from nontoxigenic strains of the organism. Most laboratories use an ELISA (enzyme-linked immunosorbent assay) method to detect the cytotoxins. The sensitivity of this test is about 85% for one stool specimen and up to 95% for 2 stool specimens (5,6,11). Therefore, the cytotoxin assay will miss 15% of cases of C.

difficile enterocolitis if one stool specimen is tested, but only misses 5% of cases if two stool specimens are tested. The specificity of this test is up to 98% (6), so false-positive results are uncommon.

Computed Tomography Computed tomography (CT) of the abdomen can reveal findings like those shown in Figure 42.2 (12). There is marked thickening in the wall of the colon, which appears as a low density region between the mucosa and serosa. The small bowel is not similarly affected. These findings are characteristic of an inflammatory process involving the large bowel (an enterocolitis), but they are not specific for C. difficile enterocolitis. Figure 42.2 Contrast-enhanced CT scan of a patient with C. difficile enterocolitis. There is marked thickening of the wall of the colon (C), but not the small bowel (SB). (From Braley SE, Groner TR, Fernandez MU, Moulton JS. Overview of diagnostic imaging in sepsis. New Horiz 1993;1:214–230.Bibliographic Links) Image digitally enhanced. View Figure


Lower GI Endoscopy Direct visualization of the lower GI mucosa is usually reserved for the few cases where the suspicion of C. difficile enterocolitis is high but the toxin assays are negative. The presence of pseudomembranes on the mucosal surface confirms the diagnosis of C. difficile enterocolitis. Colonoscopy is preferred to proctosigmoidoscopy for optimal results.

Treatment The first step in treatment is to discontinue the offending antibiotic(s) if possible, and observe for improvement. Antibiotic therapy to eradicate C. difficile is recommended only if the diarrhea is severe and associated with signs of systemic inflammation (fever, leukocytosis, etc.), or if it is not possible to discontinue the offending antibiotic(s) (13). The recommended antibiotic regimens are shown below ( 5,6,13,14). 1. Oral or intravenous metronidazole (500 mg three times daily) is the treatment of choice, and should be continued for 10 days. The oral route is preferred, and the intravenous route should be reserved for patients who are unable to receive oral medications. Both routes are equally effective. 2. Oral vancomycin (125 mg four times daily) is as effective as metronidazole, but is used as a second-line agent because of the efforts to curtail vancomycin use and

limit the spread of vancomycin-resistant enterococci. Vancomycin is preferred for pregnant or lactating females, but is otherwise reserved for cases where metronidazole is ineffective or is not tolerated. Vancomycin is not effective when given intravenously. 3. Antiperistaltic agents are contraindicated (5,14) because reduced peristalsis can prolong exposure to the cytotoxins. The expected response is loss of fever by 24 hours and resolution of diarrhea in 4 to 5 days (5). Most patients show a favorable response, and in the few who do not respond by 3 to 5 days, a switch from metronidazole to vancomycin is indicated. Although rarely necessary, surgical intervention is required when C. difficile colitis is associated with progressive sepsis and multiorgan failure, or signs of peritonitis, despite antibiotic therapy (10). The procedure of choice is subtotal colectomy. Relapses following antibiotic treatment occur in about 25% of cases (5,6,11). Most relapses are evident within 3 weeks after antibiotic therapy is completed. Repeat therapy using the same antibiotic is successful in about 75% of relapses, and another relapse is expected in about 25% of cases (14). As many as 5% of patients experience more than 6 relapses (5).

Probiotic Therapy Antibiotic therapy for C. difficile enterocolitis can itself promote persistence of the organism, and this could explain the high relapse rate. Oral administration of Saccharomyces boulardii (lyophilized yeast preparation) P.775 or Lactobacillus spp. can be used as primary prophylaxis of C. difficile enterocolitis, or can be started along with antibiotics and continued to prevent relapses of symptomatic disease. This approach has proven successful in reducing the incidence of C. difficile enterocolitis (15,16), but it has not gained favor in the United States.

Abdominal Abscess Abdominal abscesses are an important consideration in septic patients who have sustained blunt abdominal trauma or are recovering from abdominal surgery (17,18).

Clinical Features Abdominal abscesses are difficult to detect on routine clinical evaluation, as demonstrated in Table 42.1. Note that localized abdominal tenderness is present in only one third of cases, and a palpable abdominal mass is present in fewer than 10% of cases. Note also that routine abdominal films provide valuable information in fewer than 15% of patients.

Computed Tomography Computed tomography (CT) of the abdomen is the most reliable diagnostic method of detection for intra-abdominal abscesses, with a sensitivity and specificity of 90% or higher

(17,18). However, CT imaging in the early postoperative period can be misleading because collections of blood or irrigant solutions in the peritoneal cavity can be misread as an abscess. CT scans are most reliable when performed after the first postoperative week (when peritoneal fluid collections have resorbed) (18). The appearance of an abscess on an abdominal CT scan is demonstrated in Figure 42.3.

Management Immediate drainage is mandatory for all intra-abdominal abscesses (19). Precise localization with CT scanning allows many abscesses to be drained percutaneously with radiographically-directed drainage catheters. Empiric antibiotic therapy should be started while awaiting the results of abscess fluid cultures. Single drug therapy with ampicillin–sulbactam (Unasyn) or imipenem is as effective as multiple-drug regimens (20).

Urinary Tract Infections Urinary tract infection (UTI) accounts for 30% of all ICU-acquired infections, and 95% of UTIs occur in patients with indwelling urethral catheters (21). The following description is limited to UTIs in the catheterized patient. Figure 42.3 Abdominal CT scan showing a multiloculated abscess in the left upper quadrant in a post-splenectomy patient. (CT image courtesy of the Loyola University Medical Education Network,

View Figure


Pathogenesis The presence of a bladder drainage catheter in the urethra creates a 4 to 7% risk of developing a UTI per day (22). Bacterial migration along the catheter is the presumed mechanism for this risk, but there is more to the puzzle. The question that needs to be answered is why bacteria that migrate up the urethra and into the bladder are not washed out of the bladder by the urine flow. The flushing action of urine is a defense mechanism that protects the bladder from retrograde invasion by skin pathogens. This protective action explains why direct injection of bacteria into the bladder will not produce a UTI in a healthy subject (23).

Bacterial Adherence The answer to the question posed in the last paragraph is linked to observations regarding bacterial adherence to the bladder epithelium. The epithelial cells of the bladder are normally coated with Lactobacillus organisms, as shown in Figure 42.4. These organisms are not pathogenic for man, and their presence on the surface of the epithelial cells prevents organisms that are pathogenic from attaching to the bladder wall. Loss of this protective coating and subsequent colonization of the bladder mucosa with gram-negative pathogens is the critical event that eventually leads to infection in the lower urinary tract (24). This is the same phenomenon that occurs in the oral mucosa in patients who develop nosocomial pneumonias, as described in Chapter 4. The events that link urethral catheterization to bacterial adherence in the bladder are unknown. Figure 42.4 Photomicrograph showing Lactobacillus organisms blanketing a bladder epithelial cell. (From Sobel JD. Pathogenesis of urinary tract infections: host defenses. Infect Dis Clin North Am 1987;1:751–772.Bibliographic Links)

View Figure


Microbiology The common pathogens isolated from urine in medical ICU patients in the United States are listed in Table 42.2. Two surveys are included (21,25) to demonstrate that gram-negative aerobic bacilli are common isolates, but Candida albicans has emerged as a prominent isolate. Almost half (46%) of the cases of candiduria are asymptomatic (21) and therefore might not represent infection.

Diagnostic Criteria The diagnosis of urinary tract infection based on urine cultures alone is misleading in patients with indwelling urethral catheters because half of these patients will have positive urine cultures after 5 days and virtually all patients will have positive urine cultures after 30 days of urethral catheterization (26). The criteria for the diagnosis of urinary tract infection in catheterized patients are shown in Table 42.3. These criteria are from a recent consensus conference on the definition of nosocomial infections (26), and they incorporate criteria proposed by the Centers for Disease Control in 1988 (27). Note that the diagnosis of UTI requires a symptom or sign of infection (fever, etc). Unfortunately,

only fever and suprapubic tenderness apply to catheterized patients because indwelling urethral catheters eliminate the discriminating value of urgency, P.778 frequency, and dysuria. Approximately 50% of elderly patients will also develop a change in mental status in association with UTIs (28). Severe cases of urosepsis can be accompanied by multiorgan dysfunction that progresses to multiorgan failure (29). TABLE 42.2 Pathogens Isolated from the Urine of Medical ICU Patients


Incidence (% total isolates) 1990–1992*


Escherichia coli






Pseudomonas aeruginosa



Klebsiella pneumoniae









Candida albicans



*From Reference 25. † From Reference 21.

Urine Cultures The threshold for significant bacteriuria in catheterized patients is 10 5 colony forming units per mL (cfu/mL). However, colony counts as low as 10 2 cfu/mL can represent infection if growth is sustained in more than one urine sample (collected on different days) (30).

Urine Microscopy Urine microscopy is diagnostic only if an unspun urine specimen shows organisms on Gram stain or at least 3 leukocytes per high-powered field. The common practice of examining spun urine sediments has little value in identifying infection in catheterized patients.

Empiric Antimicrobial Therapy Empiric antibiotic therapy pending culture results is recommended for patients with suspected UTI who are immunocompromised, have evidence of multiorgan dysfunction, or have a prosthetic or damaged heart valve. The urine Gram stain, if positive, can be used to guide antibiotic selection. The following are some suggestions.

Gram-Negative Bacilli According to the susceptibility graphs in Chapter 44 (Figs. 44.1 and 44.2), imipenem should suffice for empiric coverage of gram-negative bacilli, and amikacin could be used for patients who have a prosthetic or damaged heart valve, or are seriously ill (i.e., are hemodynamically unstable or have evidence of multiorgan failure). TABLE 42.3 Criteria for the Diagnosis of Upper Tract UTI in Patients with Indwelling Urethral Catheters*

I. The presence of one of the following: Body temperature >38°C Urgency or frequency or dysuria Suprapubic tenderness AND II. Urine culture growing =10 5 CFU/mL of no more than 2 organisms. OR III. One of the following is present: Positive urine dipstick for leukocyte esterase or nitrates =3 WBCs per high-power field using unspun urine Organisms present on Gram stain of unspun urine 2 urine cultures with =10 2 CFU/mL of the same organism *Upper tract UTI includes infection involving the kidneys, ureters, bladder, urethra, or tissue surrounding the retroperitoneal or perinephric space. From Calandra T, Cohen J. The international sepsis forum consensus conference on definitions of infection in the intensive care unit. Crit Care Med 2005; 33:1538.


Gram-Positive Cocci A preponderance of gram positive cocci on the urine Gram's stain suggests that enterococcus is the responsible pathogen because staphylococci are uncommon offenders in nosocomial UTIs. If the patient is not seriously ill, enterococcal UTI can be treated effectively with ciprofloxacin. If the patient is seriously ill or has a prosthetic or damaged heart valve, ampicillin or vancomycin plus gentamicin is the preferred regimen. Ampicillin resistance is reported in 10 to 15% of nosocomial enterococcal infections (31), so vancomycin may be preferred. If vancomycin-resistant enterococci are a concern, linezolid can be used as a substitute for vancomycin (see the very end of Chapter 44 for a description of linezolid).

Candiduria The presence of Candida in the urine often represents colonization, but candiduria can also be a sign of disseminated candidiasis (the candiduria in this case is the result, not the cause, of the disseminated candidiasis). Disseminated candidiasis can be an elusive diagnosis because blood cultures are sterile in more than 50% of cases ( 32), and candiduria may be the only sign of disseminated disease. The following recommendations are from recent guidelines published by the Infectious Disease Society of America (33). 1. Asymptomatic candiduria in immunocompetent patients does not require antifungal therapy. However, the urinary catheter P.780 should be removed if possible because this can eradicate candiduria in 40% of cases (34). 2. Candiduria should be treated in symptomatic patients (i.e., fever or suprapubic tenderness), and patients with neutropenia or a renal allograft, because candiduria can be a sign of disseminated candidiasis in these patients. 3. Persistent candiduria in immunocompromised patients should prompt further investigation with ultrasonography or CT images of the kidney.

Antifungal Therapy Bladder irrigation with amphotericin B is not recommended because local recurrence is common (33). For non-neutropenic patients with symptomatic candiduria, fluconazole (200 to 400 mg daily) for 7 to 14 days can be effective (33). For patients with renal insufficiency or for infection with species other than Candida albicans, the new antifungal agent capsofungin (50 mg daily) is a reasonable choice (see Chapter 44 for a description of capsofungin). For all other patients, amphotericin B (0.3 to 1 mg/kg daily) for 1 to 7 days can be effective (33).

A Final Word The unifying feature in infections that involve, or originate from, the gastrointestinal, urinary, and respiratory tracts (see also Chapter 41) is the initial colonization with pathogenic organisms that first takes place. This colonization seems to involve a change

in the ability of microorganisms to adhere to epithelial surfaces. In healthy subjects, the epithelial surfaces in the mouth, GI tract, and urinary tract are covered by harmless commensal organisms, but in patients who develop an acute or chronic illness, these surfaces are covered with pathogenic organisms, and this serves as a prelude to nosocomial infections. This repopulation is not just a matter of “territorial imperative” (where one population forces another population to leave), but seems to involve the ability of microorganisms to adhere to the epithelial cells. If this is the case, then we need to study the mechanisms whereby microorganisms adhere to epithelial surfaces in health and disease if we are to effectively deal with the threat of nosocomial infections.

References Acalculous Cholecystitis 1. Walden D, Urrutia F, Soloway RD. Acute acalculous cholecystitis. J Intensive Care Med 1994;9:235–243.

2. Imhof M, Raunest J, Ohmann Ch, Rohrer H-D. Acute acalculous cholecystitis complicating trauma: A prospective sonographic study. World J Surg 1992; 1160–1166.

3. Te HS. Cholestasis in HIV-infected patients. Clin Liver Dis 2004;8:213–228. Bibliographic Links P.781

Colonization of the GI Tract 4. Marshall JC, Christou NV, Meakins JL. The gastrointestinal tract: the “un-drained abscess” of multiple organ failure. Ann Surg 1993;218:111–119. Bibliographic Links 5. Bartlett JG. Antibiotic-associated diarrhea. N Engl J Med 2002;346:334–339. Ovid Full TextBibliographic Links 6. Mylonakis E, Ryan ET, Calderwood SB. Clostridium difficile-associated diarrhea. Arch Intern Med 2001;161:525–533. Ovid Full TextFull TextBibliographic Links 7. Fekety R, Kim F H, Brown D, et al. Epidemiology of antibiotic associated colitis. Am J Med 1981;70:906–908. Bibliographic Links

8. Samore MH, Venkataraman L, DeGirolami, et al. Clinical and molecular epidemiology of sporadic and clustered cases of nosocomial Clostridium difficile diarrhea. Am J Med 1996;100:32–40. Bibliographic Links 9. Johnson S, Gerding DN, Olson MM, et al. Prospective, controlled study of vinyl glove use to interrupt Clostridium difficile nosocomial transmission. Am J Med 1990;88:137–140. Bibliographic Links 10. Lipsett PA, Samantaray DK, Tam ML, et al. Pseudomembranous colitis: a surgical disease? Surgery 1994;116:491–496.

11. Yassin SF, Young-Fadok TM, Zein NN, Pardi DS. Clostridium difficile-associated diarrhea and colitis. Mayo Clin Proc 2001;76:725–730. Bibliographic Links 12. Fishman EK, Kavuru M, Jones B, et al. Pseudomembranous colitis: CT evaluation of 26 cases. Radiology 1991;180:57–60. Bibliographic Links 13. Guerrant RL, Van Gilder T, Steiner TS, et al. Practice guidelines for the management of infectious diarrhea: Infectious Disease Society of America. Clin Infect Dis 2001;32:331–350. Full TextBibliographic Links 14. Aslam S, Hamill RJ, Musher DM. Treatment of Clostridium difficile-associated disease: Old therapies and new strategies. Lancet Infect Dis 2005;5:549–557. Bibliographic Links 15. Surawicz C. Prevention of antibiotic-associated diarrhea by Saccharomyces boulardii: A prospective study. Gastroenterology 1989;96:981–988. Bibliographic Links 16. D'Souza AL, Rajkumar C, Cooke J, Bulpitt CJ. Probiotics in prevention of antibiotic-associated diarrhea: Meta-analysis. Br Med J 2002;324:1361–1366. Bibliographic Links

Abdominal Abscesses 17. Mirvis SE, Shanmuganthan K. Trauma radiology, Part I: Computerized tomographic imaging of abdominal trauma. J Intensive Care Med 1994;9:151–163. Bibliographic Links

18. Fry DE. Noninvasive imaging tests in the diagnosis and treatment of intra-abdominal abscesses in the postoperative patient. Surg Clin North Am 1994; 74:693–709. Bibliographic Links 19. Oglevie SB, Casola G, van Sonnenberg E, et al. Percutaneous abscess drainage: Current applications for critically ill patients. J Intensive Care Med 1994; 9:191–206. Bibliographic Links 20. Mosdell DM, Morris DM, Voltura A, et al. Antibiotic treatment for surgical peritonitis. Ann Surg 1991;214:543–549. Bibliographic Links

Urinary Tract Infections 21. Richards MJ, Edwards JR, Culver D, Gaynes RP. Nosocomial infections in medical intensive care units in the United States. Crit Care Med 1999;27:887–892. Ovid Full TextBibliographic Links P.782 22. Amin M. Antibacterial prophylaxis in urology: A review. Am J Med 1992;92 (suppl 4A):114–117.

23. Howard RJ. Host defense against infection: Part 1. Curr Probl Surg 1980;27: 267–316. Bibliographic Links 24. Daifuku R, Stamm WE. Bacterial adherence to bladder uroepithelial cells in catheter-associated urinary tract infection. N Engl J Med 1986;314:1208–1213. Bibliographic Links 25. Emori TG, Gaynes RP. An overview of nosocomial infections, including the role of the microbiology laboratory. Clin Microbiol Rev 1993;6:428–442. Bibliographic Links 26. Calandra T, Cohen J, for the International Sepsis Forum Definition of Infection in the ICU Consensus Conference. Crit Care Med 2005;33:1538–1548. Ovid Full TextBibliographic Links 27. Garner JS, Jarvis WR, Emori TG, et al. CDC definitions for nosocomial infections, 1988. Am J Infect Control 1988;16:128–140.

Full TextBibliographic Links 28. McCue JD. How to manage urinary tract infections in the elderly. J Crit Illness 1996;11(suppl):S30–S40.

29. Bone RC, Larson CB. Gram-negative urinary tract infections and the development of SIRS. J Crit Illn 1996;11(suppl):S20–S29.

30. Stamm WE, Hooten TM. Management of urinary tract infection in adults. N Engl J Med 1993;329:1328–1334. Ovid Full TextBibliographic Links 31. Jenkins SG. Changing spectrum of uropathogens: implications for treating complicated UTIs. J Crit Illn 1996;11(suppl):S7–S13.

Candiduria 32. British Society for Antimicrobial Chemotherapy Working Party. Management of deep Candida infection in surgical and intensive care unit patients. Intensive Care Med 1994;20:522–528.

33. Pappas PG, Rex JH, Sobel JD, et al. Guidelines for treatment of candidiasis. Clin Infect Dis 2004;38:161–189. Bibliographic Links 34. Sobel JD, Kauffman CA, McKinsey D, et al. Candiduria: A randomized double-blind study of treatment with fluconazole or placebo. Clin Infect Dis 2000;30:19–24. Full TextBibliographic Links

Chapter 43 The Immuno-Compromised Patient When you do battle, overcome your opponent by calculation. --Sun Tzu (The Art of War) The care of critically ill patients is a labor-intensive, time-consuming, and mentally exhausting experience, and each of these aspects of patient care in the ICU reaches its zenith in the care of patients with impaired immune function. This chapter will focus on two patient populations who suffer the consequences of immune suppression: those infected with the human immunodeficiency virus (HIV), and those who develop myelosuppression from cancer chemotherapy. The care of immunocompromised patients like these is a topic of monumental size, and the material in this chapter represents only the tip of the iceberg.

The HIV-Infected Patient The introduction of highly active antiretroviral therapy (HAART) in 1996 has changed the character and outlook of patients with HIV infection who are admitted to the ICU. This is demonstrated in Figure 43.1 (1). Prior to HAART, most HIV-related admissions to the ICU were for Pneumocystis carinii pneumonia, and about 50% of the patients survived to hospital discharge (1,2). Since the introduction of HAART, the prevalence of pneumocystis pneumonia has decreased considerably, and bacterial pathogens have emerged as the most common etiologic agents in HIV-related pneumonia (1,2). Survival has also improved to the point where three of every four patients admitted to the ICU with an HIV-related disorder can now leave the hospital (1).

Figure 43.1 The change in character of HIV-related ICU admissions after the introduction of highly active antiretroviral therapy (HAART) in the mid-1990s. (Data from Reference 1)

View Figure


Pneumonia Pneumonia continues to be the most common cause of HIV-related ICU admissions. The wide spectrum of etiologic agents in HIV-related pneumonias is evident in Table 43.1 (2,3,4,5). Bacterial pneumonias are most common, while non-bacterial pneumonias occur more often in the advanced stages of HIV infection (e.g., when CD4-lymphocyte counts are below 200/mL). The shaded box in Table 43.1 highlights the observation that P.785 there is often no pathogen identified in HIV patients with suspected pneumonia (3,5). TABLE 43.1 Causes of Pneumonia in HIV Patients*

Bacterial Pneumonia

Nonbacterial Pneumonia

Streptococcus pneumoniae (15–20%)

Pneumocystis carinii (3–15%)

Hemophilus influenza (5–15%)

Fungi (5–10%)

Staphylococcus aureus (3–5%)

Mycobacterium tuberculosis and Mycobacterium avium complex (5%)

Pseudomonas aeruginosa (3–6%) Other Gram-negative bacilli (3–5%)

Organism never identified (25–60%)

Atypical organisms (1–4%)† *From References 1,2 and 3. The parentheses show the reported incidence for each microbe. † Includes Chlamydia pneumoniae, Legionella spp, and Mycoplasma pneumoniae.

Bacterial Pneumonia The most common bacterial isolates in HIV-related pneumonias are encapsulated organisms like Streptococcus pneumoniae (pneumococci), Haemophilus influenza, and Staphylococcus aureus (3,4). Pneumococcal pneumonia is the most common bacterial pneumonia, and is associated with bacteremia much more frequently than in non–HIV-infected patients. The clinical presentation and treatment of pneumococcal pneumonia is the same as in non–HIV-infected patients. There is a high rate of recurrence (10–15%) within 6 months (3), so vaccination against pneumococcal infection is particularly important in HIV-infected patients (see later). The other organism that is prevalent in HIV-related pneumonias is Haemophilus influenza. The incidence of Haemophilus influenza pneumonia is about 100-fold higher in HIV-infected patients than in the general population (3). This organism can be difficult to isolate because its growth is easily suppressed by ongoing or recent antibiotic treatment (3). The antibiotics that are effective against H. influenza include second- and third-generation cephalosporins (e.g., ceftriaxone), azithromycin, and fluoroquinolones.

Pneumocystis Pneumonia Pneumocystis carinii (renamed Pneumocystis jurovecii when it infects humans), is a protozoa-like organism (reclassified as a fungus in 1988) that proliferates in patients who are immunosuppressed. Although declining in frequency, pneumocystis pneumonia is still considered the most common opportunistic infection in HIV-infected patients ( 6), and it is almost always seen in advanced stages of the disease (e.g., when CD41 lymphocyte counts are less than 200/mL). Patients with pneumocystis pneumonia typically present with fever, non-productive cough, and hypoxemia that is out of proportion to the appearance of the chest x-ray. The initial chest x-ray can be normal in 40% of patients (7), but as the disease progresses, bilateral infiltrates like those in Figure 43.2 begin to appear. In advanced cases of pneumocystis pneumonia, the bilateral infiltrates coalesce to produce a chest x-ray like the one in Figure 43.3. The radiographic features of severe pneumocystis pneumonia are similar to those of the acute respiratory distress syndrome (ARDS) (see Fig. 22.3). The diagnosis of pneumocystis pneumonia requires visualization of the organism in specimens obtained from the respiratory tract. The most popular method of detection is the use of monoclonal antibodies directed at pneumocystis antigens (direct fluorescent antibody method), which will detect both trophic and cystic forms of the organism ( 6). Sputum induction with hypertonic saline has a diagnostic yield of 50 to 90% (3,6), although this varies with the prevalence of HIV infection (and thus the experience of cytopathologists) in individual medical centers. The highest diagnostic yield is provided by bronchoalveolar lavage, which demonstrates the organism in over 90% of cases (3). The treatment of pneumocystis pneumonia is described later in the chapter. Figure 43.2 Portable chest x-ray of an HIV-infected patient who presented with fever, nonproductive cough, and dyspnea. Routine cultures of sputum and blood were unrevealing, but bronchoalveolar lavage showed numerous Pneumocystis carinii organisms.

View Figure


Tuberculosis As many as 10% of HIV-infected patients who are purified protein derivative (PPD) positive will develop active tuberculosis (TB) each year (3). The radiographic features of pulmonary TB are determined by the CD41 lymphocyte count in blood. When the CD41 cell count is above 200/mL, the chest x-ray often shows upper lobe cavitary disease,

similar in appearance to active TB in non–HIV-infected patients. However when the CD41 cell count is below 200/mL, active TB can be accompanied by non-cavitary infiltrates in the mid-lung fields (3), and this radiographic appearance can be confused with a bacterial pneumonia.

Diagnostic Approaches to Pneumonia The clinical presentation of HIV-related pneumonia is often nonspecific, and does not allow identification of the responsible pathogen (3,6,7). In particular, the pattern of infiltration on the chest radiograph is not pathogen specific. As mentioned earlier, pneumocystis pneumonia can present with a clear chest x-ray or a non-specific radiographic pattern such as the one in P.787 Figure 43.2, and the typical appearance of advanced pneumocystis pneumonia in Figure 43.3 can also be seen in atypical pneumonias and ARDS. Identification of the etiologic agent requires blood cultures and an evaluation (with histologic examination and cultures) of sputum and bronchoscopic specimens from the lower respiratory tract. Figure 43.3 Radiographic appearance of an advanced case of pneumocystis pneumonia, which is often indistinguishable from the acute respiratory distress syndrome (ARDS).

View Figure

Bronchoscopy Bronchoscopy is a valuable diagnostic tool in HIV-related pneumonias because specimens obtained from the lower respiratory tract (by bronchial brushing or bronchoalveolar lavage) will identify over 90% of cases of pneumocystis pneumonia and pulmonary TB (3). In addition, quantitative bacterial cultures of bronchial brushings and bronchoalveolar lavage specimens can identify the responsible pathogen(s) in 70 to 80% of bacterial pneumonias (see Table 41.3). The use of bronchoscopy to diagnose bacterial pneumonias is described in Chapter 41. The ability to isolate bacterial pathogens in sputum or bronchoscopic specimens is markedly reduced in patients who are receiving antibiotic treatment. (The ability to identify Pneumocystis carinii is not affected by a few days of appropriate P.788

antibiotic coverage). Therefore, sputum and bronchoscopic specimens should be collected prior to starting antibiotic therapy, if possible.

An Organized Approach The initial approach to the HIV-infected patient with pneumonia can be guided by the CD41 lymphocyte count in blood. If the CD41 cell count is above 200/mL, then the patient should be evaluated for a bacterial pneumonia as described in Chapter 41. If the CD41 count is below 200/mL, then the management can proceed as follows: 1. Place the patient in respiratory isolation (because pulmonary TB can have a nonspecific radiographic appearance at reduced CD41 cell counts). 2. Collect sputum for Gram's stain, Ziehl-Neelson stain (for tubercle bacilli) and direct fluorescent antibody stains (for Pneumocystis carinii). Induce sputum production with nebulized hypertonic saline if necessary. Make sure sputum is screened for microscopic evidence that the specimen originates from the lower airways (see Figure 41.3). Obtain appropriate cultures (bacterial and TB) when indicated. 3. If the microscopic examination of sputum is unrevealing, perform bronchoalveolar lavage to identify Pneumocystis carinii and tubercle bacilli, and to obtain TB cultures and quantitative bacterial cultures. 4. If bronchoscopy is not immediately available, begin empiric antibiotic treatment for pneumocystis pneumonia and/or bacterial pneumonia based on clinical judgment (e.g., empiric coverage for pneumocystis pneumonia is usually given to patients with respiratory failure or diffuse infiltrates on chest x-ray). Pneumocystis can be demonstrated in the lower respiratory tract for days after starting appropriate antibiotic coverage, so bronchoscopy should be attempted, if possible, in the first few days of empiric antibiotic treatment. 5. Treatment for pulmonary TB is started only if there is evidence of infection on Ziehl-Neelson stains for tubercle bacilli. If there is no evidence of pulmonary TB on two or three sputum samples, respiratory isolation can be discontinued. Despite its diagnostic value, bronchoscopy is performed on fewer than 50% of patients with the diagnosis of pneumocystis pneumonia (8). Most of these patients are given empiric antibiotic treatment for pneumocystis pneumonia with no attempt to identify the organism. This practice should be discouraged because it mandates three weeks of (possibly unnecessary) antibiotic therapy and the antimicrobial agents are often poorly tolerated (see next).

Treatment for Pneumocystis Pneumonia Trimethoprim–sulfamethoxazole (TMP–SMX) is the antibiotic of choice for pneumocystis pneumonia. The recommended dose is 20 mg/kg of P.789 TMP and 100 mg/kg of SMX daily, administered in three or four divided doses. Although TMP–SMX can be given orally, intravenous therapy is advised for patients with respiratory failure. A favorable clinical response may not be apparent for 5 to 7 days ( 9), and there may be an initial period of deterioration. Radiographic improvement lags behind the clinical improvement (9), so the chest x-ray should not be used to evaluate the response to therapy. If a favorable response is not evident after 5 to 7 days, the treatment is considered a failure. If there is improvement in 5 to 7 days, treatment is continued for a total of 3 weeks (3,10).

Adverse reactions to TMP–SMX develop in 30 to 50% of HIV-infected patients (10,11,12,13,14). These reactions usually appear during the second week of treatment, and they are often severe enough to warrant discontinuing the drug. The most common side effects are neutropenia (45 to 50%), fever (45 to 50%), skin rash (35 to 40%), elevated hepatic transaminase enzymes (30 to 35%), hyperkalemia (30%), and thrombocytopenia (10 to 15%). A case of fatal pancreatitis has also been linked to TMP–SMX (14). Only 35 to 45% of patients who receive TMP–SMX are able to complete the full course of therapy. The high incidence of adverse reactions to TMP–SMX is specific for HIV infection. In other groups of patients, adverse reactions to TMP–SMX develop in only 10% of the patients (10). Pentamidine isothionate is the preferred second-line agent when TMP–SMX fails or is not tolerated. The recommended dose is 4 mg/kg given intravenously as a single daily dose. Intramuscular injection is not recommended because of the risk for sterile abscesses. The response time and duration of therapy are the same as for TMP–SMX. Treatment failures occur in one-third of patients (10). Adverse reactions are also common with intravenous pentamidine ( 10,15,16,17). These side effects include neutropenia (5 to 30%), hyperglycemia and hypoglycemia (10 to 30%), prolonged Q–T interval (3 to 35%), torsade de pointes (up to 20%), renal insufficiency (3 to 5%), and pancreatitis (up to 1%). Almost half of the patients who receive intravenous pentamidine are unable to complete therapy because of adverse reactions (10,15). In patients who are unable to complete therapy with either TMP–SMX or pentamidine, a variety of other agents (e.g., clindamycin and primaquine) are available. In this situation, leave the decision to an infectious disease specialist.

Steroids A brief course of steroid therapy is standard for cases of pneumocystis pneumonia that result in respiratory failure. When started at the time of antimicrobial therapy, steroid therapy is associated with improved outcomes in pneumocystis pneumonia (18). Most clinical trials used oral prednisone, but intravenous methylprednisolone (40 mg every 6 hours for at least 7 days) has also been recommended ( 10). Delay of treatment for 72 hours after the start of antimicrobial therapy negates any possible benefit from steroids (18). The response to steroids seems to vary in different clinical reports, and favorable responses can be short-lived (19).

Pneumocystis and Pneumothorax Pneumothorax is an uncommon (5% of cases) but serious complication of pneumocystis pneumonia (20). When it occurs during mechanical P.790 ventilation, as illustrated in Figure 43.4, it is usually a sign of extensive underlying tissue destruction in the lungs, and few patients survive (20). To reduce the risk of this serious complication, it seems wise to adopt the strategy known as “limited-volume” ventilation for patients with pneumocystis pneumonia. This method is designed to limit the inflation volumes during mechanical ventilation, and it is used to prevent lung injury from overdistention in patients with acute respiratory distress syndrome. A description of this method is in Chapter 22 (see Table 22.4).

Figure 43.4 Portable chest x-ray showing a pneumomediastinum and bilateral pneumothorax in a patient with pneumocystis pneumonia. The white arrows point to the outer edge of the collapsing lungs.

View Figure

Cryptococcal Meningitis Cryptococcal meningitis is the most common life-threatening fungal infection in HIV-infected patients (21,22). It is expected in 10% of patients with HIV infection, and usually appears in the advanced stages of immunosuppression (i.e., when CD4 lymphocyte counts fall below 50/mm 3).

Clinical Features The most common manifestations are fever and headache, each reported in approximately 85% of cases (21). Other findings include meningeal P.791 signs (35 to 40%), altered mental status (10 to 15%), and seizures (less than 10%) (21). Cryptococcal infections at other sites (e.g., pneumonia and skin rash) are seen in 20% of cases (22).

Diagnosis The diagnosis of cryptococcal meningitis requires lumbar puncture. Standard measurements in cerebrospinal fluid (CSF), such as glucose, protein, and leukocyte count, can be normal in up to 50% of cases (21). The organism can be demonstrated on india ink stains of CSF in 75% of cases (which is higher than the yield from india ink stains in non–HIV-infected patients) (21). CSF cultures and cryptococcal antigen titers are positive in over 90% of cases (21).

Treatment The recommended treatment for cryptococcal meningitis in HIV-infected patients is shown below (23). 1. Start with amphotericin B (0.7–1 mg/kg/day) and flucytosine (100 mg/kg/day) for

the first 2 weeks. 2. After 2 weeks, switch to oral fluconazole (400 mg/day) and continue for a minimum of 10 weeks. Thereafter, the dose of fluconazole is reduced (200 mg/day) and treatment is continued indefinitely. Chapter 44 contains a description of these antifungal agents. The mortality in this disorder is about 30% despite antifungal therapy (21).

Toxoplasmic Encephalitis Toxoplasma gondii encephalitis is the most common neurologic disorder in HIV-infected patients. Clinical evidence of toxoplasmic encephalitis is reported in 5 to 15% of HIV-infected patients, and autopsy evidence of the disease is present in up to 30% of patients (21).

Clinical Features Toxoplasmic encephalitis is characterized by focal brain lesions. Hemi-paresis and other focal neurologic deficits are seen in 60% of cases, and seizures are reported in 15 to 30% of patients (21). Other manifestations include fever (5 to 55%), confusion (60 to 65%), and choreiform movements (considered by some to be pathognomonic of toxoplasmic encephalitis) (21). Although extraneural disease is not common, disseminated toxoplasmosis with septic shock has been reported (24).

Diagnosis Computerized tomography (CT) usually reveals solitary or multiple hypo-dense, contrast-enhancing lesions in the basal ganglia and frontoparietal regions of the cerebral hemispheres (21). An example of such a lesion is shown in Figure 43.5. Note the hypodense core and the contrast enhancement at the outer edges of the lesion. Because of the radiographic P.792 appearance, these lesions are sometimes called “ring-enhancing lesions” or “ring lesions.” These lesions are not pathognomonic of toxoplasma encephalitis: similar lesions can be found in cases of lymphoma. CT scans can be unrevealing in the early stages of the disease. Magnetic resonance imaging (MRI) is more sensitive than CT scans and can reveal lesions when CT scans are negative (25). Lumbar puncture usually reveals abnormal findings, but these are nonspecific.

Figure 43.5 CT image showing a ring-enhancing lesion in a patient with toxoplasmic encephalitis. The hypodense area surrounding the lesion is evidence of cerebral edema.

View Figure

The diagnosis of toxoplasma encephalitis can be made with certainty when the organism is identified in excisional brain biopsies using immunoperoxidase staining. (Needle biopsies have a lower diagnostic yield). However, the common practice is to bypass the brain biopsy and rely instead on a presumptive diagnosis of toxoplasmosis based on the presence of characteristic lesions in the brain plus with serologic evidence of recent toxoplasma infection. Over 90% of patients with toxoplasma encephalitis will have anti-toxoplasma antibodies (IgG) in their blood, so a positive antibody titer is a sensitive marker of toxoplasma infection. Unfortunately, 20% of the general population also have these antibodies P.793 in their blood (26), so a positive antibody titer lacks specificity for toxoplasma encephalitis.

Treatment The preferred treatment for toxoplasma encephalitis is a combination of pyrimethamine (200 mg loading dose, then 75 mg daily) and clindamycin (600 mg every 6 hours). Because pyrimethamine is a folate antagonist, folinic acid (10 mg) is given with each dose of pyrimethamine to reduce the incidence of bone marrow suppression. All agents are given orally. Approximately 70% of cases show a favorable response to this regimen, and improvement is usually evident within the first week of therapy (27). The condition is considered uniformly fatal without appropriate therapy.

Drug-Related Problems The drugs currently used for antiretroviral therapy have certain adverse effects and drug interactions that deserve mention.

Lactic Acidosis Nucleoside reverse transcriptase inhibitors can be associated with a lactic acidosis caused by inhibition of mitochondrial enzymes involved in the electron transport chain.

This problem is most often associated with danosine and stavudine ( 28,29). The lactic acidosis can be severe, and a mortality of 77% has been reported (29). Case reports suggest a beneficial response to riboflavin (50 mg daily), thiamine (100 mg daily) and L-carnitine (50 mg/kg) (28), and all three can be given as treatment. The offending drug should, of course, be discontinued.

Drug Interactions Antiretroviral drugs are sometimes given to patients in the ICU. These drugs have a multitude of potential drug interactions, and the ones most likely to be seen in the ICU are included in Table 43.2. Most of these P.794 interactions are significant enough to recommend avoiding the drug combinations instead of reducing the drug dose. TABLE 43.2 Drug Interactions with Antiretroviral Drugs*





Ritonavir, Other PIs

Bradycardia, hypotension


Aprenavir, Atazanavir




Increased normeperidine



Opiate withdrawal



Enhanced sedation

*From Reference 27. PI = protease inhibitor, NRTI = nucleoside reverse transcriptase inhibitor, NNRTI = nonnucleoside reverse transcriptase inhibitor.

The Neutropenic Patient The risk of infection with neutropenia (neutrophil count less than 500/mm3) depends on the cause, severity, and duration of the neutropenia. Most cases of neutropenia that are complicated by serious infections are persistent (last longer than 10 days) and are caused by bone marrow suppression from chemotherapy (30). Neutropenia from other causes (e.g., viral infections) is rarely associated with an increased risk of infection,

particularly if the neutropenia lasts less than 10 days (30). The reason for this discrepancy is not clear, but the propensity for infections in chemotherapy-induced neutropenia may be due to additional immune suppression from the primary disease that requires the chemotherapy (i.e., cancer or organ transplantation).

Febrile Neutropenia The following statements highlight some of the important observations in patients with neutropenia and fever (30). 1. About two-thirds of patients with fever and neutropenia will not have an apparent infection on the initial evaluation. 2. Gram-positive organisms, especially coagulase-negative staphylococci, are the most frequent causes of bacterial infections in febrile neutropenia. 3. Bacteremia is relatively uncommon, and occurs in only 10 to 15% of patients with neutropenia and fever. The most likely sites of infection in neutropenic patients are the lungs, urinary tract (in patients with indwelling drainage catheters), central venous catheters, and skin (for transplant patients with surgical wounds). Whether or not there is evidence of infection at these sites, blood cultures should be obtained routinely, prior to starting empiric or directed antibiotic therapy.

Pulmonary Infiltrates The infectious and noninfectious causes of pulmonary infiltrates in neutropenic patients with fever is shown in Figure 43.6 (31). In this study, fungal pneumonia is the most common lung infection, while bacterial and viral pneumonias are uncommon. The most common cause of fungal pneumonia in this study was Aspergillus fumigatus, and the remaining isolates included Fusarium, Histoplasma capsulatum, and Candida glabrata. The bacterial isolates included staphylococci (coagulase positive and negative) and gram-negative enteric organisms, including Pseudomonas aeruginosa. The only virus isolated in this study was cytomegalovirus.

Figure 43.6 Pie chart showing the infectious and noninfectious causes of pulmonary infiltrates in neutropenic patients with fever. (Results from Piekert T, et al. Safety, diagnostic yield, and therapeutic implications of flexible bronchoscopy in patients with febrile neutropenia and pulmonary infiltrates. Mayo Clin Proc 2005;80:1414.Full TextBibliographic Links)

View Figure

P.795 The pie graph in Figure 43.6 also demonstrates that a considerable proportion of neutropenic patients with suspected fever have no evidence of infection. Forty percent of the patients in this study had no evidence of infection, and 22% had no identifiable cause for the infiltrates. The noninfectious causes of pulmonary infiltrates are listed in the figure. The most common noninfectious pulmonary disorder is diffuse alveolar hemorrhage. The prevalence of fungal pneumonias in neutropenic patients highlights the value of bronchoscopy in the diagnostic evaluation of pneumonia in these patients. Sputum is notoriously unreliable for the diagnosis of fungal pneumonia, and demonstration of the organisms deep within the lungs is required. Bronchoscopy can establish the diagnosis of fungal pneumonia with bronchial brushings, bronchoalveolar lavage, or transbronchial lung biopsy (the latter technique is not appropriate for ventilator-dependent patients or patients with thrombocytopenia). In one study that provided the data in Figure 43.6, the results of bronchoscopy resulted in a change in therapy in 50% of the patients. Bronchoscopy P.796 also provided the diagnosis in the patient whose chest x-ray is shown in Figure 43.2.

Empiric Antibiotics Prompt initiation of antibiotic therapy is recommended for all neutropenic patients with fever. This recommendation is based on the observation that patients with gram-negative septicemia due to Pseudomonas aeruginosa can deteriorate rapidly without appropriate antibiotic coverage (30). However, pseudomonas bacteremia is not common in neutropenic patients (except following kidney transplantation) (30), so the rush to antibiotics may not be justified in most patients. Regardless of timing, empiric antibiotics are recommended for all patients with neutropenia and fever (32). The choice of antibiotics is determined by the likelihood that the patient has a serious infection, as

described next.

Low Risk Patients The criteria for identifying patients who are unlikely to have a serious infection are shown in Table 43.3. The typical patient who meets these criteria has no evidence of infection, does not appear to be ill, and has neutropenia that is expected to resolve in about one week. The recommended antibiotics for such patients are shown in Table 43.4. The intravenous regimen uses one of three antibiotics: ceftazidime, cefepime, or a carbepenem (imipenem or meropenem). These antibiotics provide broad spectrum coverage, but are selected for their activity against Pseudomonas aeruginosa. I prefer imipenem because it is active against all possible gram-positive and gram-negative pathogens with the exception of methicillin-resistant Staph aureus. TABLE 43.3 Criteria For a Low Risk of Serious Infection*

History and Physical Exam Malignancy in remission. Neutropenia present 2 days)



? or NL

Primary hypothyroidism




Primary hyperthyroidism




Non-thyroidal illness: Early systemic illness

Thyroid disease:

Adapted from: Dayan CM. Interpretation of thyroid function tests. Lancet 2001; 357:619–624. Peeters RP, Debaveye Y, Fliers E, et al. Changes within the thyroid axis during critical illness. Crit Care Clin 2006; 22:41–55.

Patterns of Thyroid Function Abnormalities The changes in free T4, free T 3, and TSH levels in both thyroid disease and non-thyroidal illness are shown in Table 48.2.

Non-Thyroidal Illness Thyroid function abnormalities secondary to systemic illness (e.g., trauma or infection) occur in 70% of hospitalized patients (11,12,15). Within a few hours following the onset of illness, free T 3 is decreased in proportion to illness severity (11). With increasing illness severity, both free T3 and free T4 levels are depressed (this pattern occurs in 30 to 50% of ICU patients), and this pattern is associated with an increase in mortality (11,12). After several days of critical illness, there is a further decline in free T 3 levels, and TSH levels may be decreased (11,12). As explained earlier, TSH levels are normal in a majority of patients with non-thyroidal illness.

Thyroid Disorders

Primary thyroid disorders are characterized by changes in both free T 3 and free T4 levels (increased in hyperthyroidism and decreased in hypothyroidism) with reciprocal changes in the plasma TSH level. Hypothyroidism due to hypothalamic-pituitary dysfunction is characterized by a reduced TSH level, as explained earlier. The salient features of thyroid disorders in critically ill patients are presented next (13). TABLE 48.3 Manifestations of Thyroid Dysfunction

Hyperthyroidism Cardiovascular:

Hypothyroidism Effusions:

Sinus tachycardia

Pericardial effusion

Atrial fibrillation

Pleural effusion





Lethargy (elderly)

Skeletal muscle myopathy

Fine tremors

Elevated creatinine

Thyroid Storm:

Myxedema Coma:



Hyperdynamic shock

Dermal infiltration

Depressed consciousness

Depressed consciousness



Most cases of hyperthyroidism are due to primary thyroid disorders (e.g., Grave's disease, autoimmune thyroiditis). Chronic therapy with amiodarone, an iodine-containing antiarrhythmic agent, can also cause hyperthyroidism (16,17).

Clinical Manifestations Some of the common or characteristic manifestations of hyperthyroidism are listed in Table 48.3. It is important to note that elderly patients with hyperthyroidism may be lethargic rather than agitated (apathetic thyrotoxicosis). The combination of lethargy and unexplained atrial fibrillation is characteristic of apathetic thyrotoxicosis in the elderly (18).

Thyroid Storm An uncommon but severe form of hyperthyroidism known as thyroid storm can be precipitated by acute illness or surgery. This condition, characterized by fever, severe agitation, and high-output heart failure, can progress to hypotension and coma (19,20) and is uniformly fatal if overlooked and left untreated.

Diagnosis As shown in Table 48.2, hyperthyroidism will be accompanied by an elevated free T 4 and free T3 level, and a reduced TSH level. Because hyperthyroidism is almost always caused by primary thyroid disease, the TSH is not necessary in hyperthyroidism. P.878

Management ß-Receptor Antagonists Immediate management of troublesome tachyarrhythmias can be achieved by administering intravenous propanolol (1 mg every 5 minutes until the desired effect is achieved). Oral maintenance therapy (20 to 120 mg every 6 hours) can be used until antithyroid drug therapy is effective.

Antithyroid Drugs The two drugs used to suppress thyroxine production are methimazole and propylthiouracil (PTU). Both drugs are given orally. Methimazole is preferred to PTU because it causes a more rapid decline in serum thyroxine levels and has a lower incidence of serious side effects (agranulocytosis) (21). The initial dose of methimazole is 10 to 30 mg once a day, and the initial dose of PTU is 75 to 100 mg three times daily (19,21). The dose of both drugs is reduced by 50% after 4 to 6 weeks of therapy.

Iodide In severe cases of hyperthyroidism, iodide (which blocks thyroxine release from the thyroid gland) can be added to therapy with PTU. Iodide can be given orally as Lugol's

solution (4 drops every 12 hours) or intravenously as sodium iodide (500 to 1,000 mg every 12 hours). If the patient has an iodide allergy, lithium (300 mg orally every 8 hours) can be used as a substitute (20).

Special Concerns in Thyroid Storm In addition to the above measures, the management of thyroid storm often requires aggressive volume resuscitation to replace fluid losses from vomiting, diarrhea, and heightened insensible fluid loss. Thyroid storm can accelerate glucocorticoid metabolism and create a relative adrenal insufficiency. Therefore, in cases of thyroid storm associated with severe or refractory hypotension, hydrocortisone (300 mg IV as a loading dose, followed by 100 mg IV every 8 hours) may help correct the hypotension. Successful management of thyroid storm also requires treatment of the precipitating event ( 20,22).

Hypothyroidism Hypothyroidism is uncommon in hospitalized patients. When present, most cases represent primary hypothyroidism (23).

Clinical Manifestations Some of the more common or characteristic manifestations of hypothyroidism are listed in Table 48.3. The most common cardiovascular manifestation is pericardial effusion (24), which develops in approximately 30% of cases, and is the most common cause of an enlarged P.879 cardiac silhouette in patients with hypothyroidism (24). These effusions usually accumulate slowly and do not cause cardiac compromise. Pleural effusions are also common in hypothyroidism. The pleural and pericardial effusions are due to an increase in capillary permeability and are exudative in quality. Hypothyroidism can also be associated with hyponatremia and a skeletal muscle myopathy, with elevations in muscle enzymes (creatine phosphokinase, aldolase, lactate dehydrogenase). Enhanced release of creatinine from skeletal muscles can also raise the serum creatinine in the absence of renal dysfunction (25).

Myxedema Coma Advanced cases of hypothyroidism are accompanied by hypothermia and depressed consciousness. Although this condition is called myxedema coma, frank coma is uncommon (26). The edematous appearance in myxedema is due to intradermal accumulation of proteins (26) and does not represent accumulation of interstitial edema fluid.

Diagnosis As shown in Table 48.2, the hypothyroid patient will have a decrease in free T 3 and free T4 levels, and in primary hypothyroidism, the TSH level is elevated. A normal total serum

T4 level will virtually exclude the diagnosis of hypothyroidism.

Thyroid Replacement Therapy The treatment for mild to moderate hypothyroidism is levothyroxine, which is given orally in a single daily dose of 50 to 200 µg (27). The initial dose is usually 50 µg/day, and this is increased in 50 µg/day increments every 3 to 4 weeks. The optimal replacement dose of levothyroxine is determined by monitoring the serum TSH level. The optimal dose is the lowest dose of levothyroxine that returns the TSH to within the normal range (0.5 to 3.5 mU/L). In 90% of cases, this occurs with a levothyroxine dose of 100 to 200 µg/day (27). Oral thyroxine therapy can also be effective in severe hypothyroidism, but intravenous therapy is often recommended (at least initially) because of the risk of impaired gastrointestinal motility in severe hypothyroidism. One recommended regimen includes an initial intravenous thyroxine dose of 250 µg, followed on the next day by a dose of 100 µg, and followed thereafter by a daily dose of 50 µg (26).

T 3 Replacement Therapy Because the conversion of T4 to T3 (the active form of thyroid hormone) can be depressed in critically ill patients (26), oral therapy with T3 can be used to supplement thyroxine replacement therapy. In patients with depressed consciousness, oral T 3 can be given in a dose of 25 µg every 12 hours until the patient awakens (28). However, the benefits of T3 supplementation are unproven. P.880

A Final Word Adrenal insufficiency is considered to be common in critically ill patients, but it is difficult to determine how common it is because the rapid ACTH stimulation test is fraught with problems (e.g, which dose of ACTH to use, measuring total cortisol instead of free cortisol, etc). Patients with septic shock, severe coagulopathies, and HIV infection seem to be particularly prone to adrenal insufficiency. When a patient with any of these conditions develops unexplained hypotension or hypotension that is difficult to control with fluids and pressors, adrenal insufficency deserves consideration. You can give steroids first and then do the rapid ACTH stimulation test (don't use hydrocortisone if you are planning to do the test), or just give steroids (hydrocortisone) and see what happens. A response to hydrocortisone should be readily apparent if adrenal insufficency is a problem.

References Reviews: Adrenal Dysfunction 1. Cooper MS, Stewart PM. Corticosteroid insufficiency in acutely ill patients. N Engl J Med 2003;348:727–734.

Ovid Full TextBibliographic Links 2. Marik PE, Zaloga GP. Adrenal insufficiency in the critically ill: a new look at an old problem. Chest 2002;122:1784–1796. Full TextBibliographic Links 3. Prigent H, Maxime V, Annane D. Clinical review: corticotherapy in sepsis. Crit Care 2004;8:122–129. Bibliographic Links 4. Annane D, Sebille V, Troche G, et al. A 3-level prognostic classification in septic shock based on cortisol levels and cortisol response to corticotropin. JAMA 2000;283:1038–1045. Ovid Full TextBibliographic Links

Adrenal Insufficiency 5. Hamrahian AH, Oseni TS, Arafah BM. Measurements of serum free cortisol in critically ill patients. N Engl J Med 2004;350:1629–1638. Ovid Full TextBibliographic Links 6. Siraux V, De Backer D, Yalavatti G, et al. Relative adrenal insufficiency in patients with septic shock: comparison of low-dose and conventional corticotropin tests. Crit Care Med 2005;33:2479–2486. Ovid Full TextBibliographic Links 7. Rothwell PM, Udwadia ZF, Lawler PG. Cortisol response to corticotropin and survival in septic shock. Lancet 1991;337:582–583. Full TextBibliographic Links 8. Soni A, Pepper GM, Wyrwinski PM, et al. Adrenal insufficiency occurring during septic shock: incidence, outcome, and relationship to peripheral cytokine levels. Am J Med 1995;98:266–271. Bibliographic Links 9. Dorin RI, Kearns PJ. High output circulatory failure in acute adrenal insufficiency. Crit Care Med 1988;16:296–297. Bibliographic Links 10. Morel J, Venet C, Donati Y, et al. Adrenal axis function does not appear to be associated with hemodynamic improvement in septic shock patients systematically receiving glucocorticoid therapy. Intensive Care Med 2006;32:1184–1190. Bibliographic Links 11. Umpierrez GE. Euthyroid sick syndrome. South Med J 2002;95:506–513.

Full TextBibliographic Links P.881

Evaluation of Thyroid Function 12. Peeters RP, Debaveye Y, Fliers E, et al. Changes within the thyroid axis during critical illness. Crit Care Clin 2006;22:41–55. Bibliographic Links 13. Dayan CM. Interpretation of thyroid function tests. Lancet 2001;357:619–624. Full TextBibliographic Links 14. Burman KD, Wartofsky L. Thyroid function in the intensive care unit setting. Crit Care Clin 2001;17:43–57. Bibliographic Links 15. Fliers E, Alkemade A, Wiersinga WM. The hypothalamic-pituitary-thyroid axis in critical illness. Best Pract Res Clin Endocrinol Metab 2001;15:453–464. Bibliographic Links

Hyperthyroidism 16. Surks MI, Sievert R. Drugs and thyroid function. N Engl J Med 1995;333:1688–1694. Ovid Full TextBibliographic Links 17. Trip MD, Wiersinga W, Plomp TA. Incidence, predictability, and pathogenesis of amiodarone-induced thyrotoxicosis and hypothyroidism. Am J Med 1991; 91:507–511. Bibliographic Links 18. Klein I. Thyroid hormone and the cardiovascular system. Am J Med 1990; 88:631–637. Bibliographic Links 19. Franklyn JA. The management of hyperthyroidism. N Engl J Med 1994;330: 1731–1738. Ovid Full TextBibliographic Links 20. Migneco A, Ojetti V, Testa A, et al. Management of thyrotoxic crisis. Eur Rev Med Pharmacol Sci 2005;9:69–74. Bibliographic Links

21. Cooper DS. Hyperthyroidism. Lancet 2003;362:459–468. Full TextBibliographic Links 22. Ehrmann DA, Sarne DH. Early identification of thyroid storm and myxedema coma. Crit Illness 1988;3:111–118.

Hypothyroidism 23. Roberts CG, Ladenson PW. Hypothyroidism. Lancet 2004;363:793–803. Full TextBibliographic Links 24. Ladenson PW. Recognition and management of cardiovascular disease related to thyroid dysfunction. Am J Med 1990;88:638–641. Bibliographic Links 25. Lafayette RA, Costa ME, King AJ. Increased serum creatinine in the absence of renal failure in profound hypothyroidism. Am J Med 1994;96:298–299. Bibliographic Links 26. Myers L, Hays J. Myxedema coma. Crit Care Clin 1991;7:43–56. Bibliographic Links 27. Toft AD. Thyroxine therapy. N Engl J Med 1994;331:174–180. Ovid Full TextBibliographic Links 28. McCulloch W, Price P, Hinds CJ, et al. Effects of low dose oral triiodothyronine in myxoedema coma. Intensive Care Med 1985;11:259–262. Bibliographic Links

Critical Care Neurology There is no delusion more damaging than to get the idea in your head that you understand the functioning of your own brain. --Lewis Thomas

Chapter 49 Analgesia and Sedation Pain is a more terrible lord of mankind than even death itself. --Albert Schweitzer Contrary to popular perception, our principal function in patient care is not to save lives (since this is impossible on a consistent basis), but to relieve pain and suffering. And there is no place in the hospital that can match the pain and suffering experienced by patients in the intensive care unit. If you want an idea of how prepared we are to relieve pain and suffering in the ICU, take a look at Figure 49.1. This chapter focuses on the use of intravenous analgesics and sedatives to achieve patient comfort in the ICU. Several reviews on this topic are included at the end of the chapter (1,2,3,4,5).

Pain in the ICU Although a majority of ICU patients receive parenteral analgesics routinely (6), 50% of patients discharged from the ICU remember pain as their worst experience while in the ICU (7). This emphasizes the need for effective pain control in the ICU.

Opiophobia The problem of inadequate pain control is partly due to misconceptions about the addictive potential of opioids, and about the appropriate dose needed to relieve pain (8,9). The following statements are directed at these misconceptions. 1. Opioid use in hospitalized patients does not cause drug addiction (8).

Figure 49.1 Percentage of house staff physicians and ICU nurses who answered incorrectly when asked if diazepam (Valium) is an analgesic. (From Loper KA, et al. Paralyzed with pain: the need for education. Pain 1989;37:315.Full TextBibliographic Links) View Figure

P.886 2. The effective dose of an opioid should be determined by patient response and not by some predetermined notion of what an effective dose should be (2). Avoiding irrational fears about opioids (opiophobia) is an important step in providing adequate pain relief for your patients.

Monitoring Pain Pain is a subjective sensation that can be described in terms of intensity, duration, location, and quality (e.g., sharp, dull). Pain intensity is the parameter most often monitored because it best reflects the degree of discomfort. The intensity of pain can be recorded using a variety of scales like the ones shown in Figure 49.2. The uppermost scale (Adjective Rating Scale) uses descriptive terms, the middle scale (Numerical Ranking Scale) uses whole numbers, and the lower scale (Visual Analog Scale) records pain intensity as a discrete point placed along a line between the ends of the pain intensity spectrum. Pain intensity scales can be used to evaluate the effect of analgesic regimens in individual patients. A numerical score of 3 or less on the Numerical Rating Scale or Visual Analog Scale can be used as evidence of effective analgesia. However, it seems easier to just ask patients if their pain is well controlled. Direct communication with patients is not only the best method of determining comfort needs, it is itself a source of comfort to patients. When critically ill patients are unable to communicate directly about pain intensity, the use of surrogate signs of pain such as physiological parameters (e.g., heart rate) or elicited behaviors (e.g., facial expressions) is an unproven and probably inappropriate practice (2,10).

Figure 49.2 Three different scales for recording pain intensity. The recommended length for the numeric scales (NRS and VAS) is 10 cm. (For more information on recording pain intensity, see Hamill-Ruth RJ, Marohn ML. Evaluation of pain in the critically ill patient. Crit Care Clin 1999;15: 35.Bibliographic Links)

View Figure


Opioid Analgesia The natural chemical derivatives of opium are called opiates. Opiates and other substances that produce their effects by stimulating discrete opioid receptors in the central nervous system are called opioids. Stimulation of opioid receptors produces a variety of effects, including analgesia, sedation, euphoria, pupillary constriction, respiratory depression, bradycardia, constipation, nausea, vomiting, urinary retention, and pruritis (11). Narcotic (from the Greek narkotikos, to benumb) refers to the general class of drugs that blunt sensation and produce euphoria, stupor, or coma. Opioids are the agents most frequently used for pain relief and mild sedation in the ICU (5,6,12). They are most effective for relieving dull tonic pain, less effective for intermittent sharp pain, and relatively ineffective for neuropathic pain. Although opioids cause mild sedation, they do not cause amnesia (unless the patient goes to sleep!) (13).

Intravenous Opioids The opioids used most often in the ICU are morphine, fentanyl, and hydromorphone (5,6,12). The intravenous administration of these agents P.888 is described in Table 49.1. The doses shown in this table are the usual effective doses, but individual dose requirements can vary widely. Remember that the effective dose of an opioid is determined by each patient's response, not by the numeric value of the dose (2,16). Continued pain relief often requires continued drug administration, either as a continuous infusion or by regularly scheduled drug dosages. Intermittent, as-needed (PRN) drug administration is a recipe for inadequate pain control and is never recommended (1,2).

TABLE 49.1 Intravenous Opioid Analgesia




Loading dose

5–10 mg

1–1.5 mg

50–100 µg

Onset of action

10–20 min

5–15 min

1–2 min

Duration (after bolus)

2–3.5 hrs

2–3 hrs

30–60 min

Infusion rate*

1–5 mg/hr

0.2–0.5 mg/hr

50–350 µg/hr

demand (bolus)

0.5–3 mg

0.1–0.5 mg

15–75 µg

lockout interval

10–20 min

5–15 min

3–10 min





Lipid solubility




Active metabolites




Histamine release




Dose adjustment for




Marino - The ICU Book - 3a Edition

Related documents

1,352 Pages • 305,899 Words • PDF • 11 MB

700 Pages • 162,551 Words • PDF • 23.2 MB

52 Pages • 9,787 Words • PDF • 195.6 KB

2 Pages • 472 Words • PDF • 79.7 KB

13 Pages • 119 Words • PDF • 5.6 MB

13 Pages • 5,221 Words • PDF • 34.3 MB

72 Pages • 26,627 Words • PDF • 6.6 MB

210 Pages • 22,067 Words • PDF • 1.3 MB

125 Pages • 17 Words • PDF • 33.2 MB

187 Pages • 74,238 Words • PDF • 7.9 MB

42 Pages • 23,170 Words • PDF • 4.3 MB